the multiwavelength spectral and timing nature of the...

9HSTFMG*afcjga+

ISBN 978-952-60-5296-0 ISBN 978-952-60-5297-7 (pdf) ISSN-L 1799-4934 ISSN 1799-4934 ISSN 1799-4942 (pdf) Aalto University School of Electrical Engineering Department of Radio Science and Engineering www.aalto.fi

BUSINESS + ECONOMY ART + DESIGN + ARCHITECTURE SCIENCE + TECHNOLOGY CROSSOVER DOCTORAL DISSERTATIONS

Aalto-D

D 12

9/2

013

Extremely energetic radiation observed in the universe is attributed to the accretion of matter by compact objects: black holes and neutron stars. Microquasars, dubbed as such because of their similar phenomenological appearance to extragalactic quasars, are accreting compact objects with resolved, relativistic jets emanating from the central source. Cygnus X-3 is one of the most peculiar sources amongst microquasars. It is one of the most radio- and X-ray-luminous objects in our Galaxy, and thus a prime target for studying the emission mechanisms that arise during violent outbursts detected from the source throughout the electromagnetic spectrum. This thesis consists of constructing a hardness-intensity diagram from X-ray data with simultaneous radio observations from Cygnus X-3, searching for low-frequency quasi-periodic oscillations and using principal component analysis to distinguish the variability of emission components during outbursts.

Karri I. I. K

oljonen T

he multiw

avelength spectral and timing nature of the m

icroquasar Cygnus X

-3 A

alto U

nive

rsity

Department of Radio Science and Engineering

The multiwavelength spectral and timing nature of the microquasar Cygnus X-3

Karri I. I. Koljonen

DOCTORAL DISSERTATIONS

Aalto University publication series DOCTORAL DISSERTATIONS 129/2013



A doctoral dissertation completed for the degree of Doctor of Philosophy to be defended, with the permission of the Aalto University School of Electrical Engineering, at a public examination held at the lecture hall S1 of the school on 27 September 2013 at 12.

Aalto University School of Electrical Engineering Department of Radio Science and Engineering Metsähovi Radio Observatory

Supervising professor Prof. Martti Hallikainen Thesis advisors Doc. Diana Hannikainen Dr. Michael McCollough Preliminary examiners Prof. Thomas Maccarone, Texas Tech University, TX, USA Dr. John Tomsick, University of California, Berkeley, CA, USA Opponent Dr. Julien Malzac, Université de Toulouse, France

Aalto University publication series DOCTORAL DISSERTATIONS 129/2013 © Karri I. I. Koljonen ISBN 978-952-60-5296-0 (printed) ISBN 978-952-60-5297-7 (pdf) ISSN-L 1799-4934 ISSN 1799-4934 (printed) ISSN 1799-4942 (pdf) http://urn.fi/URN:ISBN:978-952-60-5297-7 Unigrafia Oy Helsinki 2013 Finland

Abstract Aalto University, P.O. Box 11000, FI-00076 Aalto www.aalto.fi

Author Karri I. I. Koljonen Name of the doctoral dissertation The multiwavelength spectral and timing nature of the microquasar Cygnus X-3 Publisher School of Electrical Engineering Unit Department of Radio Science and Engineering

Series Aalto University publication series DOCTORAL DISSERTATIONS 129/2013

Field of research Astrophysics

Manuscript submitted 04/03/2013 Date of the defence 27/09/2013

Permission to publish granted (date) 16/05/2013 Language English

Monograph Article dissertation (summary + original articles)

Abstract Microquasars are double star systems with a compact object, i.e. a neutron star or a black

hole, accreting matter from a companion star and occasionally channeling the accreted matter into highly relativistic particle beams called jets. Amongst microquasars Cygnus X-3 is one of the most luminous and as such is an ideal source to study the connection between gravi-tationally heated, accreted matter close to the compact object and the jet. Cygnus X-3 is also a unique source among microquasars as it contains a massive companion star that drives a strong stellar wind, which causes additional emission scenarios when interacting with the compact object and the jet. This thesis sets out to search and study the observational phenomena caused by this special set of conditions, and attempts to resolve what the link is between different emission scenarios during violent outbursts - the birth of the jet - observed from the source.

To study the disk-jet connection in detail a hardness-intensity diagram (HID) is constructed

from broadband X-ray data with simultaneous radio observations. Based on the HID a new X-ray/radio state is identified in this thesis: the hypersoft state. This state is softer in the X-ray spectra than previously classified states and exhibits very weak or nonexistent radio emission. The hypersoft state is found to be connected to the jet ejection episodes and to gamma-ray emission detected from Cygnus X-3. The fact that the sequence of events giving birth to a jet differs from other microquasars could be attributed to the strong stellar wind component and/or our line-of-sight almost coinciding with the jet axis.

In addition, quasi-periodic oscillations (QPOs) are searched for in the noise-rich X-ray light-

curves in an effort to pinpoint the environment in which these oscillations occur. The search resulted in two detections, which are associated to a certain extent with jet ejection episodes. That the QPOs coincide with outbursts points to a direct or indirect jet origin of the QPOs.

Finally, principal component analysis (PCA) is used to study the variability of X-ray emission

components during outbursts. The PCA studies demonstrate a method for eliminating spectral-fit degeneracy that the X-ray community should be enticed into embracing more widely. The PCA showed that there are two main variability components in play during outbursts. In addition, a double soft-seed population Comptonization scenario is proposed that might occur in other microquasars as well, where so far mostly single soft-seed photon Comptonization models have been used. Multiple studies have shown that such a hot component is required, supporting a different source for the seed photons besides the disk. Keywords Microquasars, Cygnus X-3, spectroscopy, time series, X-rays, radio, gamma-rays

ISBN (printed) 978-952-60-5296-0 ISBN (pdf) 978-952-60-5297-7

ISSN-L 1799-4934 ISSN (printed) 1799-4934 ISSN (pdf) 1799-4942

Location of publisher Helsinki Location of printing Helsinki Year 2013

Pages urn http://urn.fi/URN:ISBN:978-952-60-5297-7 179

Tiivistelmä Aalto-yliopisto, PL 11000, 00076 Aalto www.aalto.fi

Tekijä Karri I. I. Koljonen Väitöskirjan nimi Mikrokvasaari Cygnus X-3:n spektri- ja aikasarjaominaisuudet Julkaisija Sähkötekniikan korkeakoulu Yksikkö Radiotieteen ja -tekniikan laitos

Sarja Aalto University publication series DOCTORAL DISSERTATIONS 129/2013

Tutkimusala Astrofysiikka

Käsikirjoituksen pvm 04/03/2013 Väitöspäivä 27/09/2013

Julkaisuluvan myöntämispäivä 16/05/2013 Kieli Englanti

Monografia Yhdistelmäväitöskirja (yhteenveto-osa + erillisartikkelit)

Tiivistelmä Mikrokvasaarit ovat kaksoistähtiä, joissa tavallinen tähti luovuttaa ainetta tiheämmälle,

neutronitähdeksi tai mustaksi aukoksi luhistuneelle tähdelle. Tiheän tähden ympärille kiertyvä aine muodostaa kertymäkiekon, jossa kitka kuumentaa ainetta niin, että se muuttuu säteilyksi. Tämän lisäksi aine voi satunnaisesti kanavoitua kompaktin kohteen läheisyydestä lähes valon-nopeudella liikkuviin hiukkassuihkuihin tutkijoille vielä tuntemattoman prosessin kautta. Cygnus X-3 on yksi kirkkaimmista mikrokvasaareista. Sen kumppanitähti on raskas, voimak-kaan tähtituulen omaava tähti, eikä toista vastaavaa järjestelmää vielä tunneta galaksissamme. Tässä väitöskirjassa tutkitaan ja etsitään ilmiöitä, joita ainutlaatuiset olosuhteet aiheuttavat Cygnus X-3:ssa, varsinkin kohteesta havaittujen voimakkaiden röntgen-, radio-, ja gammapurkausten aikana.

Kertymäkiekon ja hiukkassuihkujen välistä yhteyttä tutkitaan väitöskirjassa samanaikaisilla

röntgen- ja radiohavainnoilla. Tulosten pohjalta Cygnus X-3:sta pystyttiin erottamaan uusi "hyperpehmeä" spektritila, jolloin kohteesta havaittu kova röntgen- ja radiosäteily on hyvin vähäistä tai olematonta. Hyperpehmeän spektritilan havaittiin olevan yhteydessä hiuk-kassuihku- ja gammasäteilyjaksoihin. Hiukkassuihkujen syntyminen Cygnus X-3:ssa näyttäisi tapahtuvan eri tavalla kuin muissa mikrokvasaareissa, mikä voi johtua kumppanitähden voimakkaasta tähtituulikomponentista ja/tai hiukkassuihkujen suuntautumisesta kohti Maata.

Lisäksi väitöskirjassa etsitään näennäisjaksollisia värähtelyjä kohinarikkaista röntgenvalo-

käyristä. Etsintä tuotti kaksi havaintoa, jotka molemmat voitiin yhdistää hiukkassuihkujaksoi-hin. Näennäisjaksollisten värähtelyjen samanaikaisuus purkausten kanssa viittaa vahvasti hiukkassuihkujen suoraan tai epäsuoraan osallisuuteen värähtelyjen synnyssä.

Viimeiseksi väitöskirjassa tutkitaan röntgensäteilykomponenttien vaihtelevuutta purkausten

aikana pääkomponenttianalyysin avulla. Analyysin mukaan kaksi pääkomponenttia on vastuussa lähes kaikesta röntgensäteilyn vaihtelevuudesta purkausten aikana. Analyysin pohjalta ehdotettua hahmotelmaa, jossa kaksi fotonipopulaatiota Compton-siroaa hiukkas-suihkuista voitaisiin soveltaa myös muiden mikrokvasaarien mallisovituksissa, joissa ylimääräinen fotonilähde on aiemmin osoitettu tarpeelliseksi. Pääkomponenttianalyysi osoittautui myös tavaksi vähentää aineiston sovittavien mallien määrää, minkä pitäisi olla houkutteleva menetelmä käytettäväksi yleisemmin röntgentähtitieteessä.

Avainsanat Mikrokvasaarit, Cygnus X-3, spektroskopia, aikasarja-analyysi, röntgen-, radio-, gammatähtitiede

ISBN (painettu) 978-952-60-5296-0 ISBN (pdf) 978-952-60-5297-7

ISSN-L 1799-4934 ISSN (painettu) 1799-4934 ISSN (pdf) 1799-4942

Julkaisupaikka Helsinki Painopaikka Helsinki Vuosi 2013

Sivumäärä urn http://urn.fi/URN:ISBN:978-952-60-5297-7 179

Preface

It has been quite a long journey. As much as I have explored the nooks

and crannies of its roads, many paths have been morphed by the peo-

ple along the way (some even had a map with them). I am indebted and

deeply grateful to my supervisor Diana Hannikainen after all these years

of guidance, support and friendship. This work owes also much to my

“across-the-pond” supervisor Michael McCollough without whom the sub-

stance of this thesis would have been a lot more thinner. This thesis, as

it stands, would not have been possible without your contribution. A deep

bow and a tip of the hat to you!

The research included in this thesis has been conducted in three dif-

ferent institutions: in the astronomy department of the University of

Helsinki, in the Harvard-Smithsonian Center for Astronomy and in the

Aalto University Metsähovi Radio Observatory. A huge thanks to all the

people at these institutes for effortless working experience, stimulating

discussions and mind-blowing ideas! Special thanks go to Osmi, Lauri,

Mari, Auni, Perttu, Petri, Aleksi, Ramiro, Roberto, José, Luis, Gabor,

Iouli, Alexander, Christine, Merja, Birgit, Minttu, Joni and Jonathan. A

big impact to this work has been caused by the presentations and dis-

cussions with peers and colleagues in several conferences and symposia.

Especially I would like to thank Robert Droulans for many fruitful dis-

cussions and collaboration! I am also deeply grateful to the 70+ collabo-

rators that have made the research in this thesis possible. I would like

to thank the pre-examiners Prof. Thomas Maccarone and Dr. John Tom-

sick, and the supervising professor Martti Hallikainen for their insightful

comments that made this thesis much better.

While ideas are fuel for research, there is no such thing as a free lunch. I

would like to acknowledge support from the astronomy department of the

University of Helsinki, the Smithsonian Astrophysical Observatory, the

i

Preface

Academy of Finland, Jenny and Antti Wihuri foundation, Magnus Ehrn-

rooth foundation, the Finnish Graduate School for Astronomy and Space

Physics and the Aalto University Metsähovi Radio Observatory.

I have been standing on the shoulders of giants in science as well as

in life. For the latter, I would like to thank Andreas, Ville and Aapo for

sharing the ups and downs of the life of a doctoral candidate. A huge

thanks for lifelong support to my mother Päivi and to my uncle Osmo.

Finally, a very special thanks go to my daughters Aamu and Sade, and

last but not least, to my wife Kaisa for love, life and everything!

Thank you!

Helsinki, August 19, 2013,

Karri Koljonen

ii

Preface

I have yet to see any problem, however complicated, which when you looked at

it in the right way, did not become still more complicated.

—Poul Anderson

Some morning while your eating breakfast and you need something new to think

about, though, you might want to ponder the fact that you see your kids across

the table not as they are but as they once were, about three nanoseconds ago.

—Neil deGrasse Tyson

Getting an education was a bit like a communicable sexual disease. It made you

unsuitable for a lot of jobs and then you had the urge to pass it on.

—Terry Pratchett

iii

Preface

iv

Contents

Preface i

Contents v

List of Publications vii

Author’s Contribution ix

List of Figures xi

List of Abbreviations xiii

List of Symbols xvii

1. Introduction 1

1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.1.1 X-ray binaries . . . . . . . . . . . . . . . . . . . . . . . 2

1.1.2 Microquasars . . . . . . . . . . . . . . . . . . . . . . . 5

1.1.3 Cygnus X-3 . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.2 Objective and scope . . . . . . . . . . . . . . . . . . . . . . . . 7

1.3 Dissertation structure . . . . . . . . . . . . . . . . . . . . . . 9

2. Theoretical foundation 11

2.1 Radiation mechanisms . . . . . . . . . . . . . . . . . . . . . . 11

2.1.1 Compton scattering . . . . . . . . . . . . . . . . . . . . 11

2.1.2 Synchrotron radiation . . . . . . . . . . . . . . . . . . 14

2.1.3 Thermal disk radiation . . . . . . . . . . . . . . . . . . 16

2.1.4 Continuum modifiers . . . . . . . . . . . . . . . . . . . 18

2.2 X-ray states . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.2.1 X-ray spectral states of Cygnus X-3 . . . . . . . . . . . 23

2.2.2 X-ray timing properties of Cygnus X-3 . . . . . . . . . 23

v

Contents

2.3 Relativistic jets . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.3.1 Jet ejection and major radio flares in Cygnus X-3 . . 26

2.3.2 γ-rays in Cygnus X-3 . . . . . . . . . . . . . . . . . . . 28

3. Materials and methods 31

3.1 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.1.1 Radio observatories . . . . . . . . . . . . . . . . . . . . 31

3.1.2 X-ray observatories . . . . . . . . . . . . . . . . . . . . 32

3.1.3 γ-ray observatories . . . . . . . . . . . . . . . . . . . . 34

3.2 Analysis methods . . . . . . . . . . . . . . . . . . . . . . . . . 34

3.2.1 The hardness-intensity diagram with simultaneous

radio flux density measurements as a probe to study

the disk-jet connection . . . . . . . . . . . . . . . . . . 34

3.2.2 Simulating power law noise and deriving a reliable

QPO signal with Monte-Carlo techniques . . . . . . . 37

3.2.3 Principal component analysis as a tool to unite tim-

ing and spectral information . . . . . . . . . . . . . . . 39

4. Results 43

4.1 The disk-jet connection and spectral states . . . . . . . . . . 43

4.2 Low-frequency QPOs and the jet . . . . . . . . . . . . . . . . 45

4.3 Spectral components during an outburst . . . . . . . . . . . . 45

4.4 The link between γ-ray emission and spectral states . . . . . 46

5. Discussion 49

5.1 Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

5.2 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . 54

Bibliography 57

Publications 69

vi

List of Publications

This thesis consists of an overview and of the following publications which

are referred to in the text by their Roman numerals.

I K. I. I. Koljonen, D. C. Hannikainen, M. L. McCollough, G. G. Pooley, S.

A. Trushkin. The hardness-intensity diagram of Cygnus X-3: revisiting

the radio/X-ray states. Mon. Not. R. Astron. Soc., 406, 307-319, March

2010.

II K. I. I. Koljonen, D. C. Hannikainen, M. L. McCollough. The reoccur-

rence of mHz QPOs in Cygnus X-3. Mon. Not. R. Astron. Soc., 416,

84-88, July 2011.

III K. I. I. Koljonen, D. C. Hannikainen, M. L. McCollough, R. Droulans.

2006 May-July major radio flare episodes in Cygnus X-3: spectro-timing

analysis of the X-ray data. Mon. Not. R. Astron. Soc., 429, 1173-1188,

February 2013.

IV M. Tavani, A. Bulgarelli, G. Piano, S. Sabatini, E. Striani, Y. Evange-

lista, A. Trois, G. Pooley, S. Trushkin, N. A. Nizhelskij, M. McCollough,

K. I. I. Koljonen, et al. Extreme particle acceleration in the micro-

quasar Cygnus X-3. Nature, 462, 620-623, December 2009.

V O. Vilhu, P. Hakala, D. C. Hannikainen, M. McCollough, K. Koljonen.

Orbital modulation of X-ray emission lines in Cygnus X-3. A&A, 501,

679-686, July 2009.

vii

List of Publications

VI K. I. I. Koljonen, D. C. Hannikainen, M. L. McCollough, G. G. Pooley,

S. A. Trushkin, M. Tavani, R. Droulans. The disk/jet connection in the

enigmatic microquasar Cygnus X-3. In Proc. IAU Symp. 275, Jets at All

Scales, Buenos Aires, eds. Romero G., Sunyaev R., Belloni T., Cambridge

Univ. Press, Cambridge, 24, p. 285K, February 2011.

viii

Author’s Contribution

Publication I: “The hardness-intensity diagram of Cygnus X-3:revisiting the radio/X-ray states”

The author’s main responsibility was to compile the data, conduct the

analyses, interpret the results, and write the paper.

Publication II: “The reoccurrence of mHz QPOs in Cygnus X-3”



Publication III: “2006 May-July major radio flare episodes in CygnusX-3: spectro-timing analysis of the X-ray data”



Publication IV: “Extreme particle acceleration in the microquasarCygnus X-3”

The author analyzed and discussed the X-ray data as well as the results

and commented on the manuscript.

ix

Author’s Contribution

Publication V: “Orbital modulation of X-ray emission lines in CygnusX-3”

The author provided the X-ray models used to derive the X-ray luminosi-

ties for the simulation, as well as discussed the results and commented

on the manuscript.

Publication VI: “The disk/jet connection in the enigmaticmicroquasar Cygnus X-3”



x

List of Figures

2.1 The geometry of Compton scattering . . . . . . . . . . . . . . 12

2.2 The mechanism for producing synchrotron radiation . . . . . 14

2.3 The spectrum from a classical accretion disk . . . . . . . . . 17

2.4 Line-forming processes . . . . . . . . . . . . . . . . . . . . . . 18

2.5 Canonical X-ray states from Cyg X-1 . . . . . . . . . . . . . . 20

2.6 X-ray spectral states of Cygnus X-3 . . . . . . . . . . . . . . . 24

2.7 A schematic diagram depicting the MHD acceleration and

collimation model . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.8 A Very Long Baseline Array image of Cygnus X-3 . . . . . . 27

3.1 A simplified model for the disk-jet connection in XRBs . . . . 35

3.2 A simplified, graphical example of how PCA works . . . . . . 40

4.1 Cygnus X-3 hardness-intensity diagram with the simulta-

neous radio observations . . . . . . . . . . . . . . . . . . . . . 44

4.2 The Swift/BAT ligthcurve with AGILE coverage . . . . . . . 46

5.1 The sequence of events in the 2008 multiwavelength cam-

paign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

xi

List of Figures

xii

List of Abbreviations

AGILE Astro-rivelatore Gamma a Immagini Leg-

gero

AMILA The Arcminute Microkelvin Imager Large

Array

ASM The All-Sky Monitor

BAT The Burst Alert Telescope

EXOSAT European Space Agency’s X-ray Observa-

tory

FHXR Flaring/hard X-ray

FIM Flaring/intermediate

FSXR Flaring/soft X-ray

FWHM Full width at half maximum

GBI The Green Bank Interferometer

GRID The Gamma-Ray Imaging Detector

HEASARC The High Energy Astrophysics Science

Archive Research Center

HEXTE The High Energy X-ray Timing Experi-

ment

HID Hardness-intensity diagram

HMXB High-mass X-ray binary

HS High/soft

xiii


IBIS The Imager on-Board the INTEGRAL

Satellite

IH Intermediate/hard

INTEGRAL The International Gamma-ray Astro-

physics Laboratory

IS Intermediate/soft

ISCO Innermost stable circular orbit

ISGRI The INTEGRAL Soft Gamma-Ray Imager

JEM-X The Joint European X-ray Monitor

LH Low/hard

LMXB Low-mass X-ray binary

MHD Magnetohydrodynamic

NASA The National Aeronautics and Space Ad-

ministration

PCA Principal component analysis

PCU Proportional counter unit

PDS Power density spectrum

PICsIT The Pixellated Imaging Caesium Iodide

Telescope

QPO Quasi-periodic oscillation

RRC Radiative recombination continuum

RXTE The Rossi X-ray Timing Explorer

RXTE/PCA The Proportional Counter Array

SPI The Spectrometer on INTEGRAL

UVOT The UV/Optical Telescope

WR Wolf-Rayet

xiv


XRB X-ray binary

XRT The X-ray telescope

xv


xvi

List of Symbols

a Acceleration

c Speed of light

f Fourier transform

g Electron index

h Planck constant

k Boltzmann constant

m Accretion rate

me Electron mass

mp Proton mass

n Number density

n Unit vector in the direction of photon prop-

agation

p Pressure

q Elementary charge

rg Gyroradius

s Eigenvector

t Time

v Speed

v Unit vector of velocity

y Compton y-factor

B Magnetic field strength

Bν Planck function

E Energy

F Flux

G Gravitational constant

I Intensity

K Normalization

xvii

List of Symbols

L Luminosity

M Mass

M� Solar mass

N Normal distribution

P Four-momentum

Pcomp Compton power

Pν Power density

Psynch Synchrotron power

Ptot Total power

Q Quality factor

R Radius

T Temperature

U Energy density

V Volume

α Spectral index

α0 Pitch angle

αi Incident angle of photon and electron

αf Angle between incident electron and scat-

tered photon

β Velocity in the unit of the speed of the light

γ Lorentz factor

θ Scattering angle of photon

μ0 Magnetic permeability

ν Frequency

νg Gyrofrequency

π Pi

ρ Density

σ Stefan-Boltzmann constant

σT Thomson scattering section

τ Optical depth

Γ Photon index

xviii

1. Introduction

Looking up in the sky with the naked eye we can marvel at the beauty

of the stars and planets and the Milky Way and be forgiven for thinking

of the sky as being the serene, peaceful and immutable place that hu-

mankind has imagined it to be for thousands of years before the onset

of modern astronomy. But during all that time we were missing out on

the real action. In deep space there are immensely dense and massive

objects that can unleash gargantuan amounts of energy in a very short

time. The energy radiating from these objects is such that our eyes are

not able to see it, and luckily for the well-being of the inhabitants of our

planet, Earth’s atmosphere blocks this harmful radiation from reaching

the ground. These objects are amongst the most intriguing in modern

astronomy and popular culture: black holes.

Black holes are the ultimate end point in the evolution of massive stars.

Following the implosion of a massive star, either by a supernova explo-

sion, and subsequent collapse, or a full collapse of the progenitor star, the

remnant can no longer withstand the gravitational pressure. What re-

sults is an extremely compact object: a black hole, from which not even

light can escape. If the black hole formation occurs in a double star, or

binary, system we can observe the black hole accreting matter from its

companion. We cannot see the black hole directly, but in this case we are

lucky to see the matter flowing into it. Therefore these black holes pro-

vide a unique laboratory for detailed studies of fundamental physics in

extreme gravitational fields.

Space is literally a vast, almost empty place that is sporadically dot-

ted with specks that we call galaxies. Even inside a relatively dense and

active hub in the universe such as our own Galaxy of billions of stars,

the distance between one star and the next can be measured in several

light-years. This means that most of the objects outside our Solar System

1

Introduction

appear to us essentially as point-like sources, even with the greater mag-

nifications achieved at astronomical observatories. But for astronomers

even a mere point of light contains a richness of information when viewed

across the whole range of radiation of the electromagnetic spectrum, to

which light visible to the human eye contributes only a tiny fraction.

With the appropriate instruments located on the ground, as well as in

orbit around the Earth, this point of light reveals a spectrum of different

energies that we observe changing with time. In the case of black holes

this spectrum can be treated as a puzzle, where the individual pieces are

models of the radiation processes representing the behavior of matter in

extreme gravitational fields. Solving the puzzle requires fitting the pieces

so that they ultimately complete the spectrum and are connected to each

other consistently. The point of light is therefore transposed in our minds

to an actual physical object in space, an object that can give us clues on

the basic working principles of the universe like no other laboratory that

is or could ever be capable of on Earth.

The subject of this thesis, a source called Cygnus X-3 – the third X-ray

source to be discovered in the late 1960s in the constellation of the swan,

Cygnus (Giacconi et al. 1967) – is one of the prime targets for studying

the above-mentioned extreme gravitational phenomena since it is one of

the most energetic X-ray emitters known to date in our Galaxy. We know

that Cygnus X-3 belongs to a class of objects called microquasars, a sub-

class of X-ray binaries (XRBs), where the nature of the gravitational field

near the black hole is so extreme that it launches huge outflows from the

central regions sending electrons flying in tightly collimated jets close to

the speed of light similar to their extragalactic counterparts, the quasars.

However, the true nature of Cygnus X-3 remains a mystery despite exten-

sive multiwavelength observations throughout the years.

1.1 Background

1.1.1 X-ray binaries

X-ray binaries are double star systems consisting of a compact object

(essentially either a neutron star or a black hole) and usually a non-

degenerate companion (i.e. any star that is fusing its hydrogen and he-

lium reservoir). The crux to forming an XRB is the transportation of mat-

2

Introduction

ter from the companion star to the compact object. This transportation

occurs in either of two fashions: the stars orbit each other closely enough

so that their gravitational potentials overlap thus allowing for the leak-

age of matter from the companion to the compact object or the companion’s

stellar wind powers an outflow of matter which is then partly captured by

the compact object. The technical terms for the above methods are “Roche

Lobe Overflow” and “wind accretion” respectively. The former primarily

drives the low-mass X-ray binaries (LMXBs) while the latter commonly

powers the high-mass X-ray binaries (HMXBs), although the opposite is

by no means excluded. In both cases, due to the different rotations present

in the binary (e.g. the rotation of the companion itself and the rotation

of the companion around the compact object), the matter flowing toward

the compact object possesses intrinsic angular momentum. Therefore, the

matter does not fall directly into the compact object (or onto, for a neu-

tron star primary) but instead starts to orbit it, and would orbit if forever

if not for some mechanism that transports the angular momentum away

thus causing the matter to fall ever closer to the compact object, until it

eventually falls into it. The transportation of the angular momentum cou-

pled with the rotation of the system and gravitational pull of the compact

object generates a disk-like structure around the compact object referred

to as an “accretion disk”.

Accretion disks are observed throughout the universe around a plethora

of objects from active galactic nuclei to protoplanetary systems. Apart

from the other accretion disk structures in the universe, the fact that

neutron stars and black holes are physically very small in size1 allows

the accretion disk in XRBs to attain very small radii and hence the mat-

ter acquires substantial gravitational energy before plunging into or onto

the compact object. This energy must then be radiated away before com-

ing to a full stop on the surface of a neutron star or, as in the case of black

holes, a fraction can also be advected inside the event horizon, therefore

reducing the radiation requirement. By and large, the current consen-

sus for accretion disk angular momentum transportation is in the form

of molecular viscosity and magnetic tension acting on adjacent Keplerian

1The radius of a neutron star is RNS ∼ 10 km, and the radius of a (non-rotating)black hole event horizon, often called the Schwarzschild radius, is two times thegravitational radius, RBH = 2Rg = 2GMBH/c2 ∼ 30 km MBH/(10 M�), where G isthe gravitational constant, MBH the mass of the black hole, M� the solar massand c the speed of light. For progradely rotating black holes the event horizondecreases down to one gravitational radius for a maximally spinning black hole.

3

Introduction

radii. In a given radius the Keplerian orbital velocity is slightly larger and

slightly slower than the orbital velocities of the next radius further away

and closer in from the compact object, respectively. Therefore, the adja-

cent fluid elements in the disk experience velocity shear which causes tur-

bulent viscosity between the elements extracting and dissipating gravita-

tional energy which in turn heats up the disk. The adjacent elements are

also magnetically connected (the magnetic field is frozen in the disk), and

the velocity shear in the disk causes the magnetic tension to increase be-

tween the elements which decreases the angular momentum of the inner

and increases the angular momentum of the outer element. The accretion

process turns out to be very effective at extracting gravitational energy

and according to the phenomenological model of Shakura and Sunyaev

(1973, in the literature it is often referred to as the α-disk model named

after the turbulent viscosity that in the model is parametrized as α) the

majority of the resulting thermal radiation is released at radius R ∼ 1–

10Rg, and is emitted in the form of X-rays with a mean energy around

1 kiloelectronvolt (keV2). It is due to this radiation that we can tap into

the extreme environment just outside the event horizons of black holes.

The radius furthest in where a particle can have a stable orbit, and thus

the furthest point where the inner edge of the accretion disk can reside

is called the innermost stable circular orbit (ISCO) that has the maxi-

mum orbital velocity, νISCO = 220 Hz (10 M�/MBH) for non-rotating black

holes. From this we immediately see that the X-ray emission could be

highly variable at the smallest timescales. It should be noted that other

phenomena contribute to the X-ray variability as well, namely the orbital

modulation (the companion blocking the radiation from the accretion disk)

as well as the accretion states that are discussed in Section 2.2.

XRBs have been on the observing lists of astronomers from the early

days of space-based X-ray instruments. Starting with the late 1940s, the

sky was scanned for mere minutes at a time with Geiger counters placed

inside the nozzles of sounding rockets. By the 1970s, the era of satel-

lite missions had arrived. It was indeed the rocket missions of the 1960s

that first spotted X-rays from celestial sources other than the Sun. These

anomalous X-ray objects in our Galaxy were given names according to

the constellations in which they were found: Scorpius X-1, Cygnus X-1,

Cygnus X-2 and Cygnus X-3. To this date tens of X-ray observatories

have flown and the number of XRBs has risen to about 300 sources (Liu21 kiloelectronvolt ∼ 1.6×10−16 Joules.

4

Introduction

et al. 2006, 2007). One of the most influential X-ray observatories that

has brought many observational constraints to accretion theories and to

the inner workings of XRBs has been the Rossi X-ray Timing Explorer

(RXTE). RXTE data has also been exhaustively used in this thesis for

X-ray spectral and timing studies. After the launch of RXTE at the end

of 1995 a plethora of X-ray observatories have been launch into space (in-

deed, the beginning of the millenium has been dubbed the golden-age of X-

ray astronomy), among them the International Gamma-ray Astrophysics

Laboratory (INTEGRAL) and Swift whose data is also used in this thesis.

1.1.2 Microquasars

In addition to X-ray wavelengths, XRBs have been observed at radio wave-

lengths. The scientific consensus is that this radiation arises from rela-

tivistic electrons circling around magnetic field lines and thus radiating

synchrotron radiation. The origin of this synchrotron radiation is linked

to a highly collimated outflow of plasma that has been seen emanating

from extragalactic quasars. Microquasars, dubbed as such because of

their similar phenomenological (and physical, albeit scaled down) appear-

ance to extragalactic quasars, are XRBs with resolved, relativistic jets em-

anating from the central source, that includes the accretion disk. These

jets are tightly linked to the accretion process, albeit exactly how is a very

important question and is still debated among scientists. However, the

prevailing view is that the jets are launched due to an interplay between

a strong vertical magnetic field and the rotation of the accretion disk

(Blandford and Payne 1982) and/or the black hole (Blandford and Znajek

1977), and possibly the revolving space – ergosphere – around the black

hole (Punsly and Coroniti 1990) and the accretion process itself. Thus,

microquasars provide ample opportunity to study the connection between

the inflow of matter toward the compact object and the subsequent outflow

of matter into the relativistic jets. This branch of inquiry is often referred

to as the “accretion-ejection process”, or the “disk-jet connection”, and it

not only applies to microquasars but also to quasars and protoplanetary

systems that exhibit similar structures with accretion disks and jets. Due

to their more complex structure compared to XRBs, microquasars exhibit

more varied emission processes arising from the interactions and emis-

sion of the particles in the relativistic jets. The relativistic particles in the

jets will orbit around the magnetic field lines producing synchrotron radi-

ation which peaks at radio frequencies. The very same particle population

5

Introduction

can also interact with the stellar photons of the companion or with the in-

terstellar material to produce γ-ray emission. However, the first time that

this was observed from a microquasar was in Publication IV, where γ-ray

emission was observed from Cygnus X-3. As seen in the above, it is crucial

to obtain data simultaneously throughout the electromagnetic spectrum

to understand the possible links between the different emission compo-

nents and especially the disk-jet connection. About twenty microquasars

are known in our Galaxy with the most studied being Cygnus X-1, GRS

1915+105, GX 339-4, SS 433 and Cygnus X-3.

1.1.3 Cygnus X-3

One of the most peculiar sources amongst microquasars is Cygnus X-3. It

is known for massive outbursts that emit throughout the electromagnetic

spectrum from radio to γ-rays and produce major radio flaring episodes

usually with multiple flares that peak up to 20 Jy (Waltman et al. 1995),

making it the single most radio luminous object in our Galaxy at its peak.

It is thus a prime target for studying the emission mechanisms that arise

during these violent events and probing the connection between accretion

and ejection.

The most striking feature of its X-ray lightcurve (Parsignault et al. 1972)

is a strong 4.8-hour periodicity which is attributed to the orbital modula-

tion of the binary. The same periodicity has been observed in the γ-ray

(Fermi LAT Collaboration et al. 2009) and in the infrared lightcurve (Ma-

son et al. 1986). Infrared spectral observations suggest that the mass-

donating companion in the binary is a high-mass Wolf-Rayet (WR) star

(van Kerkwijk et al. 1992). Due to these observables Cygnus X-3 is by

definition a unique source amongst HMXBs since it harbors an atypical

companion star and has a very short orbital period. On the premise that

the binary constitutes a WR companion and a black hole, this unique-

ness has been established also through population studies (Lommen et al.

2005). Similar sources have been found in other nearby galaxies IC 10

(Prestwich et al. 2007; Silverman and Filippenko 2008) and NGC 300

(Carpano et al. 2007; Crowther et al. 2010; Binder et al. 2011), where

two XRBs appear to contain WR companions and black hole primaries.

These Cygnus X-3 type sources are good candidates to being the pre-stage

in the evolution of binaries into a double black hole binary (Bulik et al.

2011).

The nature of the compact object is not certain, but several studies lean

6

Introduction

toward a black hole origin. Zdziarski et al. (2013) favor a low-mass black

hole based on orbital kinematics measured using X-ray emission lines.

Cygnus X-3 also shares spectral resemblance to other black hole XRB sys-

tems, such as GRS 1915+105 and XTE J1550−564 (see e.g. Szostek et al.

2008; Hjalmarsdotter et al. 2009) in addition to the above-mentioned ex-

tragalactic systems. Also, so far there is no evidence of a neutron star

system producing such massive radio outbursts as observed from Cygnus

X-3.

The distance to Cygnus X-3 is rather poorly constrained due to heavy

absorption in the Galactic plane that totally obscures the optical wave-

lengths. Depending on the method, the distance to Cygnus X-3 is reported

to be > 9.2 kpc (Dickey 1983), 8 kpc (Predehl and Schmitt 1995) and 9+4−2

(Predehl et al. 2000, later refined to 8.4+0.6−0.4 in Predehl et al. (2001)). The

distance measurement of Cygnus X-3 by Ling et al. (2009) is very depen-

dent upon the distance to the Cygnus OB2 association, but they give a

conservative range of 6–10 kpc. Thus, we adopt 9 kpc as the distance to

Cygnus X-3 in this thesis as it overlaps nearly all of the measurements.

The X-ray spectral and timing properties of Cygnus X-3 show a dispar-

ity contributing to the difficulty of interpreting the nature of the system.

Apart from the strong X-ray modulation, the X-ray variability of Cygnus

X-3 is otherwise very nondescript and does not show any resemblance to

other XRBs. On the other hand, the X-ray spectra of Cygnus X-3 are no-

toriously complex and more or less resemble the X-ray spectra observed

in other XRBs but differ somewhat in certain features and in the type of

spectra. This disparity could be explained by the interaction of the strong

stellar wind of the WR companion with the compact object as suggested

in Fender et al. (1999) and Szostek and Zdziarski (2008) if the accretion

disk is enshrouded partially or wholly in the stellar wind.

1.2 Objective and scope

The approach for studying Cygnus X-3 is two-fold in this thesis: spectral

information is obtained through multiwavelength observations from radio

to γ-ray wavelengths and timing information is obtained through high

time-resolution X-ray observations. The methods are complementary

to each other and help us to understand the nature of this system, and

subsequently provide us with more insight into the class of radio-jet

XRBs as a whole. The research topics of this thesis are introduced below.

7

Introduction

Topic 1: Using simultaneous X-ray and radio observations to

study the disk-jet connection in Cygnus X-3.

The classification of the different spectral states of Cygnus X-3 has

progressed on two fronts: radio (Waltman et al. 1996; McCollough et al.

1999) and X-rays (Szostek and Zdziarski 2008; Hjalmarsdotter et al.

2009). However, due to the link between accretion (producing the X-ray

emission) and ejection (producing the radio emission) in microquasars,

it seems natural to look for mutual and thus simultaneous behavior in

both emission regimes. The first step in this direction was undertaken

by Szostek et al. (2008) where X-ray and radio monitoring observations

were examined together. However, the monitoring frequencies gave only

a rough estimate of the underlying broadband X-ray energy spectral

distribution and thus a better description of the disk-jet connection might

be revealed through a more detailed study. Publication I and Publication

VI deal with this subject by utilizing simultaneous radio monitoring with

broadband X-ray spectra. The inclusion of a third wavelength regime,

γ-rays, is introduced in Publication IV to shed more light onto the disk-jet

connection and the energetics of the jet.

Topic 2: Searching for low-frequency quasi-periodic oscilla-

tions and their link to a special set of conditions in Cygnus X-3.

The X-ray timing properties of Cygnus X-3 are remarkably nondescript

apart from the strong orbital modulation. High-frequency timing phe-

nomena, such as quasi-periodic oscillations (QPOs) that are used to study

the strong gravitational regime of XRBs/microquasars, are most likely

washed out by scattering in Cygnus X-3. In addition, the most prominent

noise component is caused by red noise. There is tendency for the eye to

identify low-frequency peaks in the power density spectrum (PDS), thus

requiring careful analysis of possible low-frequency signals. However,

there are indications that there might be an increase of variability in the

form of low-frequency QPOs in the mHz regime (van der Klis and Jansen

1985). As these QPOs seem to be fleeting and sporadic events, they most

likely are caused by a special set of conditions in the system. The possible

special set of conditions was searched for in Publication II and Publica-

tion VI with the aid of radio monitoring and millisecond X-ray lightcurves.

Topic 3: Searching for the interplay of the compact object

8

Introduction

with the stellar wind in Cygnus X-3.

Cygnus X-3 displays unique properties in the X-ray spectra when

compared to other XRBs/microquasars. There are hints that these

peculiarities are linked to the interplay between the compact object and

the dense stellar wind emanating from the companion WR star (Szostek

and Zdziarski 2008). Publication III and Publication V concentrate on the

search for and the study of the effects caused by the stellar wind on the

X-ray spectra of Cygnus X-3 using a novel spectro-timing analysis method

in the former and high-resolution X-ray spectroscopy in the latter.

1.3 Dissertation structure

The overview of this thesis continues with a more detailed introduction

to radiation mechanisms, X-ray states and relativistic jets in Chapter 2.

The observations and analysis methods used in this thesis are presented

in Chapter 3. The results are presented in Chapter 4 and discussed in

Chapter 5.

9

Introduction

10

2. Theoretical foundation

2.1 Radiation mechanisms

Occasionally dubbed the “fifth form of matter”, plasma is the most im-

portant medium in the astrophysical context. The unique characteristics

of quasi-neutrality and the large mass difference between electrons and

protons give rise to a plentitude of interactions between the electrons and

photons, as well as between electrons and magnetic fields. The most im-

portant radiation processes when studying microquasars and XRBs are

Compton scattering, synchrotron radiation, thermal disk radiation and

line emission, which are briefly described in the following. A full treat-

ment of these processes can be found e.g. in Rybicki and Lightman (1979)

or Longair (2011).

2.1.1 Compton scattering

One of the most important interactions between electrons and photons in

astrophysical plasmas is Compton scattering (Fig. 2.1). In a scattering

event an incident photon with four-momentum Pγi = (Ei/c)(1, ni), where

E is the energy and n is the unit vector in the direction of propagation

of the photon, hits an incident electron moving at arbitrary speed with

four-momentum Pei = γime(c, vi), where γ = (1− β2)−1/2 = (1− v2/c2)−1/2

is the Lorentz factor, me is the mass and v is the unit vector of velocity

of the electron. This interaction results in an energy exchange for the

photon and electron leaving the scattering site with four-momenta Pγf =

(Ef/c)(1, nf ) and Pef = γfme(c, vf ), respectively. Energy-momentum con-

servation states that Pγi+Pei = Pγf +Pef which can be written as Pγf ·Pγi+

Pγf ·Pei = Pγf ·Pef . It can be shown using Lorentz invariants (PγμP γμ = 0,

PeμP eμ = −m2ec

2 and (Pγi + Pei)μ(Pγi + Pei)μ = (Pγf + Pef )μ(Pγf + Pef )μ)

11

Theoretical foundation

Figure 2.1. The geometry of Compton scattering.

that Pγf ·Pef = Pγi ·Pei, thus we get Pγf ·Pγi +Pγf ·Pei = Pγi ·Pei. Inserting

the four-momenta in vector notation and solving the above gives for the

change in photon energy

Ef

Ei=

1 − βi cos αi

1 − βi cos αf + Eiγimec2

(1 − cos θ), (2.1)

where the angles are defined as cos θ = ni · nf (the scattering angle),

vi cos αf = nf · vi and vi cos αi = ni · vi. In the electron rest frame

(βi = 0, γi = 1) the incoming radiation is symmetric, thus the average

change in energy for a photon is Ef = (1 + Ei/mec2)−1, where we see that

Ef < Ei and energy is lost to the electron recoil. However, in astrophysical

plasmas the electrons are normally highly relativistic (βi → 1, γi � 1). In

this case the photons can acquire vast amounts of energy and this process

is referred to as “inverse Compton scattering”. Qualitatively, the incoming

radiation is symmetric but the outgoing Comptonized radiation is highly

beamed in the direction of the electron velocity (cos αf ≈ 1, 〈cos αi〉 = 0),

thus we get Ef/Ei ≈ 1/(1−βi) = 2γ2i . Therefore, the photon has gained en-

ergy that is many times its initial energy. For example, with large Lorentz

factors (e.g. γ = 103) 1 GHz radio emission is boosted to 1 PHz emission,

corresponding to the ultraviolet portion of the electromagnetic spectrum;

or optical emission (ν ∼ 1014 Hz) gets boosted to γ-rays (ν ∼ 1020 Hz). Of

course, energy conservation states that the maximal energy gained by the

photon is Emax = γmec2 corresponding to the full electron energy.

12


In a low optical depth regime the Compton power is given as

Pcomp =dErad

dt=

43σT cγ2β2Urad, (2.2)

where σT is the Thomson scattering section and Urad is the initial radi-

ation energy density. From the above the total Compton power per unit

volume from a medium of relativistic electrons can be calculated provided

that the velocity distribution of electrons is known. Usually, the distribu-

tion of electrons is thermal, non-thermal or a mixture of both. For a non-

relativistic, thermal distribution of electrons 〈β2〉 = 3kTe/mec2, where k is

the Boltzmann constant Te is the temperature of the electrons, and γ ≈ 1

we get

Ptot =dPcomp

dV

∣∣∣∣∣thermal

=d2Erad

dtdV=

dUrad

dt= 4σT cUradne

(kTe

mec2

), (2.3)

where ne is the total number density of electrons. In the case of non-

thermal (γ � 1, β ≈ 1) power law electron distribution, where ne =∫Kγ−gdγ, K is the normalization of the distribution and γ ranges from

γmin to γmax the total Compton power is

Ptot =∫

Pcompne(γ)dγ =43σT cUrad

K

3 − g(γ3−g

max − γ3−gmin ). (2.4)

Qualitatively, the resulting energy spectrum in the power law case (re-

calling Ef ∝ γ2Ei from above) is

d2Erad

dtdV dEf=

dγ

dEf

d2Erad

dtdV dγ∝ 1

γγ2−g = γ1−g ∝ E

(1−g)/2f . (2.5)

Thus, the scattered energy spectrum is a power law but with a spectral

index α = (g − 1)/2 and a photon index Γ = (g + 1)/2.

In the discussion so far we have considered only a single scattering

event. However, several scattering events can take place depending on

the electron scattering optical depth τes ≈ neσT R, where R is the size

of the scattering region in question. For multiple scattering events in-

verse Compton scattering is characterized by a Compton y-factor, which

is defined as the average fractional energy gained per scattering by the

photons multiplied by the mean number of scattering events. The mean

number of scattering events is obtained from the optical depth τes, where

a value of τes ∼ 1 means that on average, a photon will scatter once before

escaping the region. Specifically, the mean number of scattering events

is approximately max(τes, τ2es) and in the case of non-relativistic, ther-

mal Comptonization the fractional energy per scattering is dUrad/Urad =

(4kTe/mec2)dτes. Thus, the y-factor is defined as

y ≡ 4kTe

mec2max(τes, τ

2es). (2.6)

13


B

Figure 2.2. The mechanism for producing synchrotron radiation. A charged particle spi-raling around a magnetic field line is accelerated and emits radiation into atight radiation cone.

The slope of the scattered radiation energy spectrum is characterized by

the y-factor so that when y 1 there is negligible change in the en-

ergy spectrum, whereas y ∼ 1 is governed by the Kompaneets equation

and describes a power law energy spectrum with an exponential cut-off.

In this case the power law portion of the energy spectrum has a slope

α = −(3/2) ± √(9/4) + (4/y) (Rybicki and Lightman 1979), where the

‘+’ correspond to the up-scattered and the ‘−’ correspond to the down-

scattered photons. Thus, it is possible to obtain a scattered power law

spectrum, even if the scattering electrons have a thermal distribution.

The values y � 1 are called saturated Comptonization where the incident

photon spectrum approaches the distribution of the scattering electrons,

i.e. the electrons and photons will come into thermal equilibrium where

the scattered photon spectrum will peak at E � 2.8kTe.

2.1.2 Synchrotron radiation

Synchrotron radiation arises from the relativistic motion of charged par-

ticles in a magnetic field, and thus is a prominent radiation component in

astrophysical plasmas since strong magnetic fields are present in many

astrophysical sources. The origin of this radiation is usually due to elec-

trons moving at relativistic speeds, since they are lighter than protons

14


and hence more easily accelerated. The non-relativistic case is referred

to as “cyclotron radiation”. The motion of electrons in the presence of a

magnetic field consists of a constant velocity component along a magnetic

field line and circular motion around it, i.e. a spiral path with a constant

pitch angle (Fig. 2.2). The radius is usually referred to as the “gyroradius”

(rg) of the electron and its corresponding frequency the “gyrofrequency”:

νg =(β⊥/rg)

2π=

qB

2πγme, (2.7)

where B is the magnetic field strength and q is the elementary charge.

Thus, the parallel and perpendicular components of acceleration to the

direction of magnetic field are (a‖ = 0) and a⊥ = νgβ⊥, respectively. The

total emitted power of the particle can be obtained via the relativistic

Larmor formula: (μ0q2/6πc)γ4(γ2a2

‖ + a2⊥), where μ0 is the magnetic per-

meability of free space. Inserting a⊥ and integrating over all pitch angles

(〈sin2 α0〉 = 2/3) for a given velocity β gives for the synchrotron power:

Psynch =23

μ0

6πc

q4

m2γ2β2B2. (2.8)

A more familiar form is achieved when the above equation is expressed

in terms of the Thomson cross-section σT = μ20q

4/6πm2c2 and magnetic

energy density UB = B2/2μ0:

Psynch =43σT cγ2β2UB, (2.9)

which is exactly the same as for Compton scattering power with the

change of the radiation energy density Urad to magnetic energy density

UB. Synchrotron radiation can be thought to arise from Compton scat-

tering off of virtual photons of the magnetic field. Thus, an important

relation in the astrophysics is

Pcomp

Psynch=

Urad

UB, (2.10)

which provides a self-regulating mechanism for a synchrotron-emitting

plasma: high-energy electrons produce photons that in turn get inverse

Compton scattered by the very same electrons, thus removing energy, or

cooling, the electrons. In the literature this interplay is known as the

synchrotron self-Compton mechanism. If the radiation energy density is

greater than the magnetic energy density in the beginning, this mecha-

nism eventually extracts all the available energy from the electrons and

shuts down the synchrotron emission. Also, similar to Compton scatter-

ing, the photon spectrum arising from the synchrotron emission of elec-

trons with a power law distribution is a power law with a spectral index of

15


α = (g−1)/2, where the g is the original index of the electron distribution.

At low frequencies the optical depth for the synchrotron photons increases

and they are unable to escape the region, producing a low energy cut-off

at ν ∝ γ2 and a spectrum that follows a power law with a spectral index

of −5/2.

2.1.3 Thermal disk radiation

In order to cool the relativistic electrons via Compton scattering a popu-

lation of soft photons is needed. One possible source of soft photons is the

above-mentioned synchrotron radiation, but in the case of microquasars

soft photons can also arise from blackbody radiation of thermal plasma in

the accretion disk and from bremsstrahlung radiation, i.e. the radiation

emitted by electrons when accelerated by the Coulomb field of another

charged particle, although the latter is usually considered to be a negli-

gible radiation process in microquasars. However, it might play a role in

the spectral energy distribution of Cygnus X-3 as discussed in Publication

III.

If the radiation field has uniform temperature it emits a characteristic

blackbody spectrum, whose intensity is given by the Planck function:

Bν =2hν3

c2

1ehν/kT − 1

, (2.11)

where h is the Planck constant. Integrating over all frequencies reduces

the above equation to B = (σ/π)T 4, where σ is the Stefan-Boltzmann con-

stant, thus the radiation of a blackbody emitter is governed by its tem-

perature alone. In the accretion disk the temperature of the plasma is

governed by the distance from the central source. For an accretion rate

m the kinetic energy gained from the gravitational potential is turned

into heat at a rate (1/2)mv2, which gives the accretion luminosity for the

non-relativistic case:

L =12

12mv2 =

GMm

2R=

12mc2 Rg

2R, (2.12)

where the first half comes from the virial theorem that states that only

half of the gravitational potential energy can be radiated. Eq. 2.12 shows

that for constant accretion rate the luminosity of the disk is inversely pro-

portional to the distance (i.e. L ∝ R−1). On the other hand, if this lumi-

nosity is produced by blackbody radiation the luminosity is proportional to

the temperature (i.e. B ∝ T 4 ⇒ L ∝ R2T 4). Thus, the temperature profile

of the disk is proportional to the radius as T (R) ∝ R(−3/4). A full analysis

16


Figure 2.3. The spectrum from a classical accretion disk where black body radiation fromdifferent annuli ranging from the ISCO (Rin) to the disk outer radius (Rout)is summed up to form a multi-color accretion disk model.

gives for the temperature profile for the non-relativistic case (Shakura &

Sunyaev, 1973):

T (R) =

(3GMm

8πσR3

[1 −

(1 − Rin

R

)1/2])1/4

. (2.13)

This is usually referred to as the “classical” disk, or optically thick, ge-

ometrically thin disk and it is used in this thesis as the model for the

accretion disk emission. However, many problems remain in modeling

the accretion disk accurately. To name a few, the inclusion of metallicities

in the plasma and the advection of matter through the event horizon of a

black hole will modify the emission spectrum. For high luminosities the

thin disk approximation breaks down and a lot of effort has been made

to study optically thin and geometrically thick disks. In particular, the

high luminosity of the plasma near the compact object can drive a large

wind off the outer portion of the disk, thus varying the m which was con-

sidered a constant in the classical solution. However, observations are

inconclusive as to which solution is operating and when.

The energy spectrum, or intensity, of the classical disk (Fig. 2.3) can

be obtained by inserting Eq. 2.13 into Eq. 2.11 and integrating over the

whole disk from its inner edge Rin to the outer edge Rout:

Iν ∝∫ Rout

Rin

2πRBν [T (R)]dR. (2.14)

17


Figure 2.4. Line-forming processes in the stellar wind near the ionizing X-ray radiationof the compact object.

Using the relation T ∝ R−3/4 and changing the variables appropriately to

dR ∝ (1/T )1/3d(1/T ) this changes to:

Iν ∝∫ 1/Tout

1/Tin

(1T

)4/3

ν3 1ehν/kT − 1

(1T

)1/3

d

(1T

). (2.15)

Simplifying this by using a variable x = hν/kT gives:

Iν ∝ ν3

ν8/3

∫ hν/Tout

hν/Tin

x5/3 1ex − 1

dx. (2.16)

In the middle of the spectrum we can approximate the limits as hν/Tin 1 ∼ 0 and hν/T out � 1 ∼ ∞, which reduces the integral to a constant

value. Thus, the spectrum goes as Iν ∝ ν1/3 in the frequency range cor-

responding to a range in annuli between Rin and Rout. The frequencies

below ν = kTout/h are in the Rayleigh-Jeans portion of the spectrum and

thus Iν ∝ ν2. Respectively, the frequencies above ν = kTin/h are in the

Wien portion of the spectrum and thus Iν ∝ exp(−hν/kTin).

2.1.4 Continuum modifiers

On top of the continuum produced by the processes described above the

final spectrum received by the detector gets modified by absorption and

18


subsequent line emission (see Fig. 2.4 for a collection of line-forming pro-

cesses). Especially in HMXBs, such as Cygnus X-3, the massive stellar

wind emanating from the companion star brings matter into the vicinity

of the intense radiation field of the compact object. In the X-ray band

one of the cooling processes through which photons lose energy to the

surrounding matter is photoionization. In this process an incident pho-

ton hits an ion or an atom and ejects an electron whose binding energy

is less than the energy of the photon. The energy levels equivalent to

the binding energy for the electrons inside atoms are called absorption

edges, as photons with higher energy than the binding energy will be ab-

sorbed by these atoms. This process produces a sharp edge in the energy

spectrum that reduces the number of photons received above the binding

energy and scales roughly as E−3. The inverse process of photoioniza-

tion where an ion captures an electron is called radiative recombination.

As the captured electron energies are distributed continuously the result-

ing emission from radiative recombination forms a continuum (radiative

recombination continuum or RRC). As this process is dependent on the

characteristics of the electron population it can be used as a diagnostic

tool of the plasma. The RRCs can be found in the energy spectrum above

the recombination edge with a width that is approximately the same as

the temperature of the electrons kTe. The ions can capture the electrons

to any free energy level, and if it is not the ground state it leaves the ions

in an excited state, followed by the subsequent decay of the excited state

to the ground state releasing spontaneous emission. This emission gen-

erates photons that will contribute to the emission line of the ions in the

plasma. Another process contributing to the emission (and absorption)

line formation is photoexcitation, where a photon excites a bound state of

an ion resulting in emission when the excited electron decays back to the

original energy level.

Another important process for line emission from XRBs is fluorescence,

where an inner-shell electron of ions in a low charge state (i.e. “cold” ions)

are photoionized leaving the ions in an excited state. This is followed by

the filling of the hole of the ejected electron with an electron from higher

energy levels which produces a photon. Alternatively, this energy can be

transferred to another electron, called an Auger electron, which is then

ejected from the ion. Perhaps the single, most important emission line

in XRBs is the fluorescence Kα emission line for neutral iron at 6.4 keV,

which is produced close to the compact object and can be used to study

19


Figure 2.5. Canonical X-ray states from Cyg X-1 (low/hard state: blue, high/soft state:red). Modified from McConnell et al. (2002)

the plasma in the vicinity of the strong gravitational field of the compact

object.

In addition to the modifications to the energy spectrum close to the com-

pact object the radiation gets modified by the interstellar matter along

the line-of-sight that produces an extra absorption factor spanning from

the infrared to the soft X-rays.

2.2 X-ray states

Soon after the discovery of XRBs they were found to be highly variable in

the X-ray band from 1 to 10 keV (Tananbaum et al. 1972), varying from

a low luminosity state to a high luminosity state. Often these states are

referred to as “the low state” and “the high state” (e.g. van der Klis 1994;

Nowak 1995; McClintock and Remillard 2006). However, one has to bear

in mind that these labels were assigned in order to explain the source be-

havior in this relatively narrow band. If a wider bandwidth is taken into

account it is possible that the low state is more luminous, i.e. the bolomet-

ric luminosity of the source is greater in the low state than the bolometric

luminosity in the high state. As mentioned in Section 1.1.1 the α-disk

model was established to explain the blackbody-like high state spectra.

In the high state the accretion disk dominates the geometry of the system

20


with little or no hot corona, and most of the bolometric luminosity is con-

centrated on photon energies below 10 keV resulting in a “soft” spectrum.

However, the low state is dominated by a power law-like component which

extends beyond 10 keV (a weaker high energy tail is seen in the high state

as well). This component is taken to represent the inverse Comptoniza-

tion of soft seed photons in a plasma cloud of hot electrons. These hot

electrons can often be located in multiple places as described below. Thus,

they are often dubbed just as “the corona”, a term which embodies all the

different possibilities for the origin of the hot electrons. In the low state

the coronal component in the source dominates, and the disk is either

greatly reduced or then immersed inside the corona resulting in a “hard”

spectrum where the bolometric luminosity is concentrated on photon en-

ergies above 10 keV. This spectral bimodality has been added to the state

vocabulary so that the high state becomes the high/soft (HS) state and

the low state the low/hard (LH) state. These states will be referred to as

the canonical X-ray states of XRBs in this thesis (see Fig. 2.5). In the

“intermediate” states, when the source is transiting from the HS state to

the LH state or vice versa, we see the influence of both components (disk

and corona) to differing extents.

The common view is that there is a change in the rate at which matter

flows (m) from the companion star to the accretion disk that ultimately

drives the spectral evolution. An important limit is set when consider-

ing the radiation pressure gradient from the intense X-ray radiation near

the ISCO and the gravitational pressure gradient of the inflowing mat-

ter. The limit where the radiation pressure gradient becomes equal to the

gravitational pressure gradient, therefore stalling the accretion of matter

toward the compact object, is called the Eddington luminosity or Edding-

ton limit, and it was originally derived for stars (spherically symmetric

objects), but it can be used to approximate the behavior of accretion disks

as well. Equating these two pressure gradients yields:

dp

dR= −GM(R)ρ(R)

R2. (2.17)

When taking into account only the Thomson scattering of photons from

the electrons of the inflow the radiation pressure gradient is proportional

to the luminosity L of the disk so that

dp

dR= − σT ρL

4πR2mpc. (2.18)

21


Inserting this into Eq. 2.17 gives the Eddington luminosity:

LEdd =4πGMmpc

σT� 1.3 × 1038 M

M�erg/s. (2.19)

When the accretion rate is high (close to the Eddington luminosity) the

material in the accretion disk pushes close to the ISCO and radiates pre-

dominantly thermal, blackbody radiation driving the system into the HS

state by cooling the hot electrons via Coulomb collisions and Compton

scattering. On the other hand, when the accretion rate is low the inflow

toward the compact object is much reduced and the Coulomb coupling be-

tween protons and electrons is weak. This in turn leads to diminished

radiation and drives the system into the LH state, as most of the heat

is contained in protons, which do not radiate, and energy transfer to the

electrons is inefficient. The hot flow is then cooled by advecting the mat-

ter into the black hole, thus the designation of this scenario: advection-

dominated accretion flow (Ichimaru 1977; Narayan and Yi 1994). There is

still an ongoing debate as to whether the accretion disk is truncated away

from the ISCO (e.g. Tomsick et al. 2009) or if the accretion disk radiation

close to the ISCO is just highly diminished (Reis et al. 2009). Another

scenario favored for producing the high energy tail in the HS state is usu-

ally attributed to active regions above the disk that are caused by small

patches of energetic electrons energized by a magnetic reconnection of the

magnetic field lines of the accretion disk. Since microquasars contain rel-

ativistic and thus Comptonizing electrons in the jet, it can also be added

to the list of possible locations for Comptonizing the disk photons. There-

fore, the radio emission – indicating the presence of the jet – is also an

important measure of the state classification. This is particularly impor-

tant for strong radio sources like Cyg X-3 as will be described in Section

2.3.1.

In addition to the canonical states XRBs exhibit various transition or in-

termediate states in between the HS and LH states. These intermediate

states are by virtue much more rare and transitionary but they convey im-

portant information on the relationship between the two canonical spec-

tral components in XRBs. One of the most widely used tools of the trade

for studying the spectral evolution and intermediate states is a hardness-

intensity diagram (HID, described in more detail in Section 3.2.1), which

shows the X-ray intensity (i.e. source luminosity not corrected for dis-

tance) as a function of hardness ratio (a ratio between the intensities of

two spectral bands, usually one for soft X-rays with intensities below 10

keV and one for hard X-rays with intensities above 10 keV). In spite of

22


the fact that XRB X-ray lightcurves are highly variable, the hardness ra-

tio is much more stable and usually exhibits smoother changes and it is

therefore a good proxy of the accretion state of the system.

2.2.1 X-ray spectral states of Cygnus X-3

Cygnus X-3 displays multiple X-ray states, including not only the canoni-

cal HS and LH states but also three (Szostek et al. 2004) or four (Publica-

tion I) other states in between (see Figure 2.6). The classification of these

states is defined by their spectral hardness and the strength of the soft,

disk-like component of the spectra. The first successful physical interpre-

tation of the broadband X-ray spectrum (Vilhu et al. 2003) consisted of a

thermal Comptonization component and Compton reflection from an ion-

ized medium with parameters similar to those found for black hole bina-

ries at high Eddington rates, with the exception of very strong absorption.

Ever since, all the broadband X-ray spectra from all the spectral states of

Cygnus X-3 have been fitted with a model where the main spectral compo-

nent is thermal/non-thermal Comptonization (Hjalmarsdotter et al. 2009,

Publication I). Thus, it appears that Cygnus X-3 is Comptonization domi-

nated. The peculiar cut-off at ∼ 30 keV in the harder spectra, which is not

observed in any other XRBs, can be explained by Compton downscattering

of high energy photons by relatively cooler electrons that are part of the

plasma cloud surrounding the compact object (Zdziarski et al. 2010). This

plasma cloud could form as the result of an interaction between the com-

pact object and the stellar wind from the WR companion. The interaction

of the Comptonized photons with the stellar wind is evident in the form

of strong emission lines, particularly iron lines, that are formed through

photoionization. Thus, the unique environment of Cygnus X-3 produces

unique spectral features.

2.2.2 X-ray timing properties of Cygnus X-3

The PDS of Cygnus X-3 was studied in Willingale et al. (1985) and was

found to be well described by a power law of index α = 1.8 in the frequency

range 10−5–0.1 Hz. This result has been more or less verified in further

studies by Choudhury et al. (2004, α ∼ 1.5), Axelsson et al. (2009, α ∼ 2

in the canonical HS state and α ∼ 1.8 in the LH state). van der Klis and

Jansen (1985) found short intervals (5–40 cycles) of mHz quasi-periodic

features in the PDS during the HS state and possibly also during the LH

23


10 100

0.010

0.100

1.000

1: Quiescent2: Transition3: FHXR4: FIM5: FSXR6: Hypersoft

keV

2 Pho

tons

s−1

cm−2

keV

−1

6543

2

1

Energy [keV]

Figure 2.6. X-ray spectral states of Cygnus X-3. The spectra are colored and labeledwith different colors and numbers according to the legend. The state nomen-clature follows the one defined in Publication I, where the quiescent andtransition states correspond to the LH X-ray state with moderate radio emis-sion, the flaring/hard X-ray (FHXR), the flaring/intermediate (FIM) and theflaring/soft X-ray (FSXR) states correspond to the intermediate X-ray stateswith radio flaring, and the hypersoft state corresponds to the HS state withquenched radio emission.

24


state using data from the European Space Agency’s X-ray Observatory

(EXOSAT). A mHz feature in the Cygnus X-3 hard X-ray lightcurve was

also found in a balloon flight study by Rao et al. (1991). On the contrary,

no QPOs were found in the RXTE study in Axelsson et al. (2009) in either

the HS or LH states of the source. The discrepancies between the reported

results clearly show that these transient features are short-lived and spo-

radic or time-sensitive events. In addition, no strict method of calculating

the significance of the peaks in the PDS was undertaken in the previous

studies.

2.3 Relativistic jets

The twin jets are very highly collimated and as they propagate through

the interstellar medium at relativistic speeds, they harbor a huge amount

of kinetic energy.

There are two important observations which tell us that the jets could

emanate from deep within the potential well of the accreting compact ob-

ject and, therefore, are closely related to the accretion process itself. First,

the speed of the jet is of the order of the escape velocity from the vicinity

of the compact object. For example, the jets observed from the black hole

candidate GRS 1915+105 move with a speed of 0.92c (Mirabel and Ro-

dríguez 1994) which corresponds to the escape velocity at a distance of

2.4Rg from the compact object. Secondly, the jet ejection takes place dur-

ing the state transition from the LH state to the HS state, and indicates a

relationship between the jet and the accretion disk. This is often accom-

panied by the temporary disappearance of the hard X-ray emission which

indicates the cessation of hard X-ray production in the corona.

Though it is clear that the black hole systems produce energetic and rel-

ativistic jets, they present a difficulty for the electromagnetic theory of jet

production: black holes cannot support a magnetic field. The only way for

supporting such a magnetic field is that the current generating the mag-

netic flux comes from an external supply of plasma, i.e. the accretion flow.

On the other hand, accretion flows are believed to be weakly-magnetized

plasmas in highly turbulent, rotating flows around the black hole (Balbus

and Hawley 1998). Thus, managing a well-behaving, global and rotating

magnetosphere with this turbulent accretion disk is a tricky problem. The

leading model for producing this kind of phenomenon is called the magne-

tohydrodynamic (MHD) accelerator model which was originally proposed

25


by Lovelace (1976) and Blandford (1976). Briefly, it uses energy from the

differential rotation of the accretion disk coupled with a strong electro-

magnetic field to convert rotational kinetic energy into kinetic energy of a

jet. A magnetic pressure gradient lifts the plasma out of the gravitational

potential well and the pinch effect collimates the outflow into a jet. How-

ever, there are observational reasons for believing that the same source

may produce jets of different speeds, either simultaneously or in different

accretion states. Similarly, there are multiple theories for producing jet

launching in the same black hole system. Two models that may be driv-

ing XRBs are known as the Blandford-Payne (Blandford and Payne 1982)

and the Blandford-Znajek (Blandford and Znajek 1977) mechanisms. The

former uses only Newtonian mechanics to drive axially-oriented outflows

from the disk and the latter concerns the vicinity of the compact object

and how in this region the magnetic field lines are forced, by the rotation

of the disk, to spin. The spinning field lines will assume an outward spi-

raling shape and centrifugal forces will fling out the magnetically frozen

plasma along the field lines to form two magnetized jets. The toroidal

magnetic field will pinch the plasma towards the jet axis that provides

self-collimation to the jets. It is also possible to extract energy from the

spin of the black hole to drive the jet and achieve more energetic outflows

(Punsly and Coroniti 1990). Supporting observations exists that show

that the jet energies in the black hole systems are indeed ∼ 30 times more

strong than the jet energies in the neutron star systems with similar ac-

cretion rates (Migliari et al. 2003).

2.3.1 Jet ejection and major radio flares in Cygnus X-3

Unlike most other XRBs, Cygnus X-3 is relatively bright in the radio vir-

tually all of the time and it undergoes giant radio outbursts with strong

evidence of precessing jet-like structures (Fig. 2.8) moving away at rel-

ativistic speeds (v ∼ 0.6–0.8c) with an inclination of ≤14◦ to the line-of-

sight (Mioduszewski et al. 2001; Miller-Jones et al. 2004). Tudose et al.

(2010) have shown that during the active states the radio emission from

the jet dominates that from the core. Also, the radio emission does not

show any orbital modulation therefore being consistent with the notion

that the radio emission arises further out in the jet. Cygnus X-3 ex-

hibits a power law of positive or negative slope in the radio, usually de-

picted as synchrotron radiation from the relativistic electrons in the jet.

This synchrotron radiation may extend to infrared/optical wavelengths

26


Figure 2.7. A schematic diagram depicting the MHD acceleration and collimation model.Magnetized and rotating inflow spiraling toward a compact object (whitesolid arrow) winds the magnetic field lines into a rotating helical coil. Mag-netocentrifugal forces expel some of the material along the field lines andmagnetic pressure and pinching forces (black short solid arrows) further liftand collimate it into a jet outflow (black long arrows).

Figure 2.8. A Very Long Baseline Array images of Cygnus X-3, modified from Mio-duszewski et al. (2001), showing a curved jet.

27


and even to X-rays and it could be optically thick or thin to synchrotron

self-absorption.

The radio emission exhibits several correlations with the X-ray emission

during a major flare episode. During times of moderate radio brightness

(∼ 100 mJy), and low variability in the radio and X-ray fluxes, the hard

X-ray flux anti-correlates with the radio (McCollough et al. 1999). The

major flaring episode begins when Cygnus X-3 descends into a radio/hard

X-ray quenched state (denoted as the hypersoft state in Publication I),

where the hard X-rays and radio switch from an anti-correlation to a cor-

relation. This state will last from a couple of days up to a month. It is a

rare state, comprising ∼ 2–3 % of the overall monitoring live time. Upon

emerging from this state, Cygnus X-3 will always exhibit major radio flar-

ing (Szostek et al. 2008) followed by flaring in the hard X-rays as well

(McCollough et al. 1999). After the major flare decay Cygnus X-3 returns

back to the LH state or it can loop again back to quenched state to produce

another major flare.

2.3.2 γ-rays in Cygnus X-3

Cygnus X-3 has been detected by various space missions out to energies

of ∼ 200–300 keV. Throughout the 1970s and 1980s there were reported

detections of Cygnus X-3 in the MeV to PeV energy ranges (Vladimirsky

et al. 1973; Danaher et al. 1981; Lamb et al. 1982; Samorski and Stamm

1983; Bhat et al. 1986), but none of these detections were confirmed or

found to be reproducible (O’Flaherty et al. 1992; Hermsen et al. 1987; Mori

et al. 1997; Aleksic et al. 2010). Recent observations by Astro-rivelatore

Gamma a Immagini Leggero (AGILE) and Fermi have shown that the

Cygnus region is a very complicated area at γ-ray energies. Cygnus X-3

has now been shown by both AGILE (Publication IV) and Fermi (Fermi

LAT Collaboration et al. 2009) to be a γ-ray source (> 100 MeV). Also,

Fermi observations show that the γ-ray emission exhibits the same 4.8

hour modulation as in X-rays and infrared attributed to orbital modu-

lation, thus resulting in the firm association of γ-rays with Cygnus X-3.

Interestingly, the peak of the γ-ray emission occurs near phase 0.0 at the

far side of the WR star. The peak isotropic luminosity above 100 MeV is

Lγ ∼ 1036 erg/s.

γ-rays have also been detected outside of major radio flare episodes and

have been associated with very brief periods of hard X-ray quenching

and minor radio flaring (Corbel et al. 2012). In one case the hard X-ray

28


quenching lasted only ∼ 0.5 day and in the other for a day. In the last

period of activity a 1 Jy radio flare (at 15 GHz) was observed during the

period of γ-ray emission (Bulgarelli et al. 2012). γ-rays are reported to

occur also without associated radio emission Williams et al. (2011).

It is still under debate whether the γ-rays arise from leptonic (Dubus

et al. 2010; Zdziarski et al. 2012b; Piano et al. 2012) or hadronic pro-

cesses (e.g. Romero et al. 2003; Piano et al. 2012). The leptonic scenario

is based on the inverse-Compton scattering of soft stellar photons by the

relativistic electrons in the jet that are streaming towards us close to the

line-of-sight. The interaction of the jet with the stellar photon field close

to the compact object (within ten orbital separations) scatters the photons

up to MeV energies. The hadronic scenario, where the jet is populated

with protons, is based on the proton-proton collisions that occur between

the protons in the hadronic jet and the protons in the stellar wind. The

collisions produce pions and γ-rays via the decay of neutral pions. In addi-

tion, the jet origin of the γ-rays is still debatable as Williams et al. (2011)

present some evidence against the model of inverse-Compton scattering

of soft stellar photons by the jet.

In this chapter the mechanisms responsible for producing the emission

from Cygnus X-3 throughout the electromagnetic spectrum have been

briefly introduced and how they connect to previous observations. In the

next chapter the observatories used to obtain the data analyzed in this

thesis are introduced, as well as the analysis methods used to extract in-

formation from the data.

29


30

3. Materials and methods

3.1 Observations

A large set of multiwavelength observations of Cygnus X-3 from ground-

and space-based observatories is used in this thesis. The main focus lies in

the analysis of the radio and X-ray data, but supporting observations from

ultraviolet/optical, infrared and γ-ray data are taken into an account. In

the following, the main instruments involved are introduced.

3.1.1 Radio observatories

Before its decommissioning in 2000, the Green Bank Interferometer

(GBI) included three 26-meter-diameter radio telescopes operated by the

National Radio Astronomy Observatory with support from the National

Aeronautics and Space Administration (NASA) High Energy Astrophysics

program at its Green Bank site in West Virginia, United States. The tele-

scope had an angular resolution of about 3 arcseconds and 11 arcseconds

in its two operating frequency bands, 8.3 GHz and 2.25 GHz, respectively.

The root-mean-square noise in a 5-minute scan is about 6 mJy at 8.3 GHz

and 10 mJy at 2.25 GHz. The data used in this thesis were obtained via

ftp://ftp.gb.nrao.edu/pub/fghigo/gbidata/gdata/2030+407.

The Arcminute Microkelvin Imager Large Array (AMILA), op-

erated by the Cavendish Astrophysics Group, is composed of eight

12.8-meter-diameter, equatorially mounted Cassegrain antennas, which

were previously part of the Ryle Telescope (decommissioned in 2006)

located at the Mullard Radio Astronomy Observatory in Lord’s Bridge

near Cambridge, England. The antennas are separated by distances

ranging between 18 and 110 m. The telescope has an angular resolution

31

Materials and methods

of approximately 30 arcseconds with a flux sensitivity 3 mJy s−1/2 and

it operates at six frequencies between 13.9–18.2 GHz (AMI Consortium:

Zwart et al. 2008). 15 GHz monitoring data of Cygnus X-3 is used in this

thesis (Publication I; Publication IV).

The RATAN-600 (or the Academy of Science Radio Telescope) is a

reflector-type radio telescope operated by the Special Astrophysical

Observatory of the Russian Academy of Science located near the village of

Nizhny Arkhyz, Russia. The main mirror of the telescope was built in the

shape of a ring with a diameter of 577 meters, containing 895 individual

elements with dimensions of 2 meters times 11.4 meters. The individual

elements reflect the radiation either toward a central conical receiver or

to one of five cylindric reflectors, that in turn direct the converging beam

to the secondary mirror, which collects the radiation at the focus where

the primary receiver feeds are located. The RATAN-600 has a maximum

angular resolution of 2 arcseconds with a flux density limit of 0.5 mJy.

It operates in the frequency range 0.61–30 GHz. Monitoring data from

Cygnus X-3 in seven frequency bands (1.0, 2.3, 4.8, 7.7, 11.2, 21.7 and 30

GHz) is used in this thesis (Trushkin et al. 2008, Publication I).

3.1.2 X-ray observatories

The Rossi X-ray Timing Explorer (RXTE) was an X-ray satellite

operated by NASA/Goddard Space Flight Center, decommissioned in

2012, consisting of three X-ray instruments: the All-Sky Monitor (ASM,

Levine et al. 1996), the Proportional Counter Array (RXTE/PCA, Jahoda

et al. 1996) and the High Energy X-ray Timing Experiment (HEXTE,

Rothschild et al. 1998). The ASM scanned about 80% of the sky every

orbit at a spectral range 2–10 keV down to a sensitivity of about 10

mCrab. The RXTE/PCA and HEXTE were pointing instruments which

cover spectral ranges 2–60 keV and 15–250 keV with an energy resolution

below 18% at 6 keV and 15% at 60 keV respectively. The RXTE/PCA

consisted of five xenon gas proportional counter units (PCU) with a

throughput of ∼ 2500 counts/s/PCU when observing the Crab Nebula,

background of 90 mCrab and time resolution 1 μs. The HEXTE consisted

of two clusters of scintillation detectors with a sensitivity of ∼1 mCrab

for 100 ks observation, background 50 counts/s/cluster and time sampling

rate 8 μs. The RXTE data used in this thesis were obtained from the

High Energy Astrophysics Science Archive Research Center (HEASARC).

32


Swift (Gehrels et al. 2004) is a space-based multiwavelength obser-

vatory operated by NASA/Goddard Space Flight Center that includes

three instruments onboard: the Burst Alert Telescope (BAT) operating in

the 15–150 keV hard X-ray band, the X-ray Telescope (XRT) operating in

the 0.3–10 keV soft X-ray band, and the UV/Optical Telescope (UVOT)

operating in the wavelength range 170–650 nm. The BAT is a coded

aperture mask instrument that has a large field-of-view of 1.4 steradians.

It is used to monitor the hard X-ray sky with a sensitivity of ∼ 2 mCrab

in 16 hours exposure time and with a spectral resolution of ∼7 keV. The

XRT is a focusing X-ray telescope with 110 cm2 effective area, 23.6 ×23.6 arcminutes field-of-view and 18 arcsecond resolution. The spectral

resolution at launch was ∼ 140 eV at 6 keV. The UVOT is a diffraction-

limited 30 cm Ritchey-Chrétien reflector, sensitive to magnitude 22.3 in

a 17 minute exposure in B filter. The Swift data used in this thesis were

obtained from the HEASARC.

The International Gamma-Ray Astrophysics Laboratory (IN-

TEGRAL) is an international space-based multiwavelength observatory

in a highly eccentric orbit designed to observe the sky in the optical, X-ray

and γ-ray wavelengths. It is operated by the European Space Agency

in cooperation with the Russian Space Agency and NASA. INTEGRAL

hosts four instruments onboard: the Imager on-Board the INTEGRAL

Satellite (IBIS, Ubertini et al. 2003) that observes from 15 keV to 10

MeV, the Spectrometer on INTEGRAL (SPI, Vedrenne et al. 2003) that

observes from 20 keV to 8 MeV, the Joint European X-ray Monitor

(JEM-X, Lund et al. 2003) that observes from 3 keV to 35 keV, and the

Optical Monitoring Camera (Mas-Hesse et al. 2003) sensitive to radiation

in the wavelength range 500 to 580 nm. IBIS is a coded-mask instrument

featuring two detector layers: the hard X-ray detecting layer INTEGRAL

Soft Gamma-Ray Imager (ISGRI, Lebrun et al. 2003) and the γ-ray

detecting layer Pixellated Imaging Caesium Iodide Telescope (PICsIT, Di

Cocco et al. 2003). IBIS has 12 arcminutes angular resolution with 8.3

× 8.0 degree fully coded field of view, time sampling rate of 61 μs and a

spectral resolution of 8% at 100 keV (ISGRI) and 10% at 1 MeV (PICsIT).

SPI is a coded-mask instrument with 2.5 degree angular resolution, 16

degree fully coded field of view and energy resolution of 2.2 keV at 1.33

MeV. Like IBIS and SPI, JEM-X is also a coded-mask instrument. It has

33


4.8 degree fully coded field of view, 3 arcminute angular resolution, 1 ms

time sampling rate and a spectral resolution of 1.3 keV at 10 keV.

3.1.3 γ-ray observatories

Astro-rivelatore Gamma a Immagini Leggero (AGILE) is a multi-

wavelength observatory operated by the Italian Space Agency that in-

cludes two instruments: the Gamma-Ray Imaging Detector (GRID, Bar-

biellini et al. 2002; Prest et al. 2003) sensitive to photons with energies

30 MeV – 30 GeV with the optimal range from 0.1–1 GeV, and the Super-

AGILE detector (Feroci et al. 2007) sensitive to photons with energies 18–

60 keV. The GRID allows γ-ray source positioning with error box radius

near 5–20 arcminutes and it is designed to achieve a nominal spectral res-

olution of δE/E ∼ 1 at 200 MeV. AGILE operates in a fixed-pointing mode,

implying that it can accumulate data on a source within its large field of

view (2.5 steradians for the GRID) fourteen times a day, taking into ac-

count Earth occultations during each spacecraft orbit. AGILE monitoring

observations of Cygnus X-3 are used in this thesis in Publication IV.

3.2 Analysis methods

3.2.1 The hardness-intensity diagram with simultaneous radioflux density measurements as a probe to study the disk-jetconnection

The “disk-jet connection” concept consists of studying the relation be-

tween the accretion disk (emitting mostly in the X-rays) and the jet (emit-

ting mostly in the radio). In general, many bright XRBs are accompanied

during an X-ray outburst by a radio outburst (Fender et al. 2009). Thus,

is is apparent that rapid changes in the disk configuration and mass ac-

cretion rate that cause X-ray state changes are also coupled with the jet

production. In the case of persistent XRBs, like Cyg X-1, the X-ray and

radio are correlated during the LH state (e.g. Zdziarski et al. 2011). But

when the system transitions from the LH to the HS state, the radio emis-

sion quenches abruptly. However, there is evidence that an unresolved

compact jet is present in the HS state at least for Cyg X-1 (Rushton et al.

2012). Again, the state transition is linked to the major changes in the

accretion geometry which prevents the jet production (e.g. Tigelaar et al.

34


Figure 3.1. A simplified model for the disk-jet connection in XRBs showing an X-ray HIDof GX 339−4 (Belloni et al. 2005) together with sketches of the accretionstates (marked with roman numerals) depicting the components of the sys-tem: the jet, the accretion disk and the corona. In the HID the logarithmof X-ray hardness ratio is plotted against the logarithm of X-ray bolometricintensity (3–200 keV). The spectral evolution of GX 339-4 (other XRBs showvery similar behavior) is shown to progress from the LH state (right), throughthe intermediate states (IH/IS) to the HS state (left) and descending back tothe LH state. In the LH and IH states the jet is steady with an almost con-stant bulk Lorentz factor typically of γ < 2, progressing from the accretionstate i to state ii as the intensity increases. During these states the diskis cold and the coronal emission is dominating the X-ray spectrum. Whenthe system crosses the so called “jet line”, Lorentz factor of the jet rapidly in-creases producing an internal shock in the outflow (iii) followed in general bycessation of jet production (albeit some fading optically thin radio emissioncan be detected from the jet/shock, which is now physically decoupled fromthe central engine). It it possible for XRB to loop around the jet line multipletimes producing further optically thin radio outburst but only when crossingthe line from right to left. At the same time the source spectrum starts to bedominated by the disk emission and eventually it moves to the HS state withhot disk, no jet and very little, if any, coronal emission (iv).

35


2004). Thus, it is the “intermediate state”, often subdivided into multiple

(sub-)states, that conveys important information regarding the disk-jet

connection, since it lies between the two extreme states. Typically, XRBs

follow a similar evolution through the different X-ray spectral states and

trace similar tracks in the HID (Fig. 3.1). A typical black hole HID rep-

resenting the transient outbursting cycle has a Q-type shape that can be

divided into four main areas or states: quiescent, low/hard, intermediate

and high/soft. These areas/states are briefly reviewed in the following:

Quiescent state (lower right in the diagram): The source flux is very low

in both the X-ray and radio with the spectrum described by a hard power

law;

Low/hard (LH, right side): The source can be bright in the X-rays and

its spectrum is described by a power law with exponential cut-off at ∼ 100

keV. The radio emission has a flat spectrum and appears as a compact jet;

Intermediate (top): As the source moves from the LH state to the in-

termediate state the radio emission stays steady almost until the point

of major ejection in the intermediate/soft (IS) state. However, the radio

emission starts to become more variable exhibiting a peaked or optically

thin spectrum shortly before the radio flare, which occurs during the state

change from the intermediate/hard (IH) state to the IS state. Also, when

the source moves in the opposite direction, from the IS to the IH, this

often corresponds to optically thin radio outbursts (Fender et al. 2004).

Jets during the intermediate states are associated with highly relativistic

motion, unlike the steady jets in the LH state which are usually mildly

relativistic. This implies that the geometrical change in the disk could

cause a more energetic jet than it would otherwise. One scenario to ex-

plain this behavior invokes internal shocks. The increase of the X-ray

luminosity of the accretion source could result in the increase of the ve-

locity of the outflow. Thus, the more energetic jet (with greater velocity)

catches up with ejecta in the jet that were emitted in the LH/IH state

and forms shocks. During this state there is a prominent thermal disk

component and a steep power law without a cut-off;

High/soft (HS, left side): This state is thermally dominated and typ-

ically there is no power law component, or if present, it is very weak.

There is also a line that demarcates the intermediate state region from

the HS region that is referred to as the jet line. When the source crosses

this line going from right to left there is a sudden cut-off of radio emission.

It is during this transition that strong radio flares occur.

36


The overall picture emerging from the disk-jet connection and theoreti-

cal studies of jet launching (see Section 2.3) is that in the LH state (phase

i in Fig. 3.1) the jet is powered by the Blandford-Payne mechanism, i.e.

the accretion disk powered jet, resulting in a steady but rather weak jet.

On the other hand, as the system transits to intermediate X-ray states

the accretion disk propagates inward and the plasma starts to “feel” the

compact object and the jet gains more energy via the Blandford-Znajek

mechanism. When the disk approaches the ISCO, the Lorentz factor of

the jet increases rapidly (see Fig. 3.1) before the jet is suppressed in the

HS state. This may occur due to the suppression of the jet launching

mechanism or due to a suppression of the radio emission. An example of

the former could be a shift in the accretion geometry from a thick disk to

a thin disk, which would strongly affect the magnetic field strength driv-

ing the jet. An example of the latter could be the interaction of the non-

thermal, synchrotron-emitting particles with the strong radiation field of

the disk slowing down the relativistic particles and producing thermal

jets. Thus, major radio flaring periods are of prime importance for study-

ing the relationship between the accretion disk and the radio jets, using

simultaneous radio and X-ray observations.

3.2.2 Simulating power law noise and deriving a reliable QPOsignal with Monte-Carlo techniques

X-ray QPOs are one of the prime observation tools for probing the strong

gravitational field near the compact object in XRBs. They range from a

low frequency regime of mHz variations to a high frequency regime of

kHz variations bounded by the light crossing time of the emission region.

The rapid variations in the X-ray lightcurve are essentially stochastic pro-

cesses and can be described best with statistical tools. This variability is

usually in the form of a power law in the Fourier transformed frequency

space spanning a large continuous set of frequencies Pν(ν) ∝ ν−α, where

Pν is the power density associated with a frequency ν. Often this broad

structure is referred to as the “noise” component. In addition to the noise,

the PDS can contain a peak around a specific frequency ν0 that can be

modeled with a Lorenzian profile Pν ∝ FWHM/[(ν − ν20) + (FWHM/2)2],

where FWHM is the full width at half maximum of the peak. The

Lorentzian in the power density spectrum corresponds to an exponentially

damped sinusoid in the lightcurve (although the underlying signal may be

something different as well), thus being a quasi-periodic oscillation with

37


the usual convention that the quality factor Q ≡ ν0/FWHM which mea-

sures the coherence of the peak is larger than two (van der Klis 2006).

A noise component with Pν(ν) ∝ ν−2 power spectrum is often referred

to as red noise. Red noise is known to produce low-frequency peaks in the

PDS and thus any low-frequency QPO candidate found in the red noise

dominated PDS has to be considered carefully and should be compared to

properties of randomly generated lightcurves (Benlloch et al. 2001).

One good method is to produce randomly generated lightcurves with

similar characteristics as the original lightcurve, i.e. the same length

and mean counts of the lightcurve, and power law index of the PDS de-

rived from the lightcurve. On the basis of a large number of simulated

lightcurves, confidence limits can be placed for each frequency correspond-

ing to how many times the simulated lightcurves reached a given power

density in a given frequency. This method is by definition a Monte-Carlo

calculation. Typically, any peak below a confidence level of 99% or, prefer-

ably, 99.9% is considered as being part of the noise (Bevington and Robin-

son 1992).

One algorithm for generating simulated power law noise lightcurves is

depicted in Timmer and Koenig (1995). It is based on a linear theory of

stochastic processes, where the Fourier transform f(ν) is a complex gaus-

sian random variable determined by the spectral density distribution:

f(ν) = N [0,12P (ν)] + iN [0,

12P (ν)], (3.1)

with a variance that does not depend on the number of data points. The

individual steps for the algorithm goes as follows:

• Choose a spectrum and take the square root of√

P (ν)

• Draw two Gaussian-distributed random number sets with length equal

to the number of frequencies in the spectrum, zero mean and standard

deviation equal to one and multiply them with√

P (ν)

• Allocate the Fourier transform f(ν) with positive and negative frequen-

cies equal to the number of frequencies in the spectrum

• Populate the Fourier transform f(ν) real and imaginary part for positive

frequencies with the random number sets drawn above

38


• Make sure that the Fourier transform is real at the Nyquist frequency

f(νNyquist) = 0.0

• Populate the Fourier transform real and imaginary part for negative

frequencies using f(−ν) = f∗(ν), where the asterisk denotes complex

conjugation

• Collect all parts (real and imaginary) to form the Fourier transform f(ν)

• Make transform into time domain using fast Fourier techniques and

drop the imaginary part

• Normalize lightcurve to match the original one

3.2.3 Principal component analysis as a tool to unite timing andspectral information

The main use of principal component analysis (PCA) is to find the smallest

number of components that suffice to explain the data losing the least

amount of information in the process. Basically, PCA finds patterns in

the data in a way that highlights the differences and similarities in the

data set. In many-dimensional cases, when graphical representation is

not convenient, it is a particularly powerful analysis tool. By finding the

“new coordinates” of the data set (i.e. the principal components) where

the data points mainly cluster and ignoring the small scatter that the

data points have around these coordinates the dimensionality of the data

set is reduced greatly, defined only by these new coordinates.

Here, we briefly summarize the main points of the analysis using X-ray

spectra as input and relating that to a simplified example depicted in Fig.

3.2:

• Start with a number of spectra p measured at times t1, t2, . . . tp binned

into n bins corresponding to energies E1, E2, . . . En. In Fig. 3.2 one data

point would correspond to one energy spectrum. However, in this case

the data points have just two energy bands (x and y), while usually the

spectra would include at least ten different bands to get a feeling of the

shape of the spectrum. The number of data points in the example would

39


Figure 3.2. A simplified, graphical example of how the PCA works. The principal com-ponents can be though of as a new coordinate system with the origin at themean data value and the axes rotated so that the first principal componentpoints to a direction where most of the data are oriented. By definition theprincipal components are orthogonal to each other. In order to reduce thedegrees of freedom (i.e. noise) from the data, the data points can be projectedon to a smaller set of principal components. E.g. in the figure the data pointsare projected to the line defined by the first principal component.

correspond to the number of spectra p.

• Form a p × n matrix with coefficients F (tp, En), i.e. the fluxes in each

energy band at times p. In the example this would be a matrix with

two columns for the values x and y and the number of rows equal to the

number of data points.

• Subtract a mean flux F (En) from the coefficients for each energy chan-

nel.

• Compute the n × n covariance matrix out of the data matrix, which

states the variances (in the diagonal) and covariances (elsewhere) be-

tween each energy bands.

• Calculate the eigenvectors, sn, and eigenvalues of the covariance ma-

trix. These eigenvectors form the new coordinates of the data and the

accompanying eigenvalue states the proportion of variance of a particu-

lar eigenvector, i.e. the highest eigenvalue and accompanying eigenvec-

tor is the first principal component of the data set etc. In the example

40


these would refer to the new coordinates PC1 and PC2.

• Form a linear decomposition Snk = (sn1sn2 . . . snk) of the data set us-

ing the above eigenvectors, ordered by the proportion of variance. The

components producing only a small fraction of the overall variance can

be then dropped off, reducing the dimensionality, i.e. choosing a small

number of eigenvectors for the linear composition. In the example the

data shows biggest variance in the direction of PC1.

• The data set defined by the new coordinates can be obtained by

Fkp = Skn × Fnp =n∑

i=1

SkiFip (3.2)

Where the chosen k coordinates lie in direction of the calculated eigen-

vectors sk. In the example this can be visualized by rotating the original

coordinates and transferring the origin to the mean of x and y so that

they correspond to the coordinates of PC1 and PC2.

• The original data set without the less significant components, which

most likely are just systematic errors and noise, can be obtained via

FPCAnp = (Snk × Fkp) + F (En) = F (En) +

k∑i=1

SniFip (3.3)

leaving only the intrinsic variability in the data set. In the example this

refers to projecting the data points onto the PC1 axis.

One of the most interesting avenues to follow in the subsequent analysis

of the principal components is to track the evolution of Fkp for each k (that

produces most of the overall variance) with time tp. Identifying this evolu-

tion with the change of spectral fit parameters performed on the original

energy spectra can give us clues as to which spectral component or a pa-

rameter of that component could be the source of the variability attributed

to the principal component snk for chosen k:s.

In this chapter the observatories used to obtain the data in this thesis

were introduced together with the three most important methods to ana-

lyze this data. The results of these analyses are presented in the following

chapter.

41


42

4. Results

4.1 The disk-jet connection and spectral states

In Publication I, a more complete HID is constructed with the help of

an extensive data set from RXTE. In addition, during most of the RXTE

observations there are (near-)simultaneous radio observations from the

GBI, AMILA and RATAN-600 radio telescopes. In the spirit of the orig-

inal HID design where the jet emission phases are depicted (connected

to radio emission via synchrotron radiation), Publication I presents the

HID of Cygnus X-3 with the simultaneous radio flux densities plotted on

top of the X-ray hardness-intensity plot (Fig. 4.1). In this way the con-

nection between the X-ray emission, both soft (from the accretion disk or

from some other thermal emission process) and hard X-rays (from inverse-

Compton scattering), and radio emission from the synchrotron process in

the jet become clearer. The radio emission demarcates the HID into three

separate regions: LH X-ray state with quiescent radio emission levels (∼100 mJy), intermediate X-ray state with radio flaring (0.3–20 Jy), and

HS X-ray state with quenched radio emission. Due to the complexity of

Cygnus X-3’s X-ray spectra, they have been classified into many different

substates in previous studies (Szostek et al. 2004; Szostek and Zdziarski

2008; Hjalmarsdotter et al. 2009). However, the region in the HID that

encompasses the HS X-ray state with the quenched radio emission does

not correspond well to any of the previously determined states as it ex-

hibits even softer X-ray spectra than the softest state in Szostek et al.

(2004); Szostek and Zdziarski (2008) called this the ultrasoft state. In

addition, the behavior of Cygnus X-3 between the softest state, dubbed

the “hypersoft” state in Publication I, and the intermediate/flaring state

exhibits very interesting phenomena. This region corresponds to the jet

43

Results

Figure 4.1. Cygnus X-3 hardness-intensity diagram with the simultaneous radio fluxdensities (the darker color referring to lower flux) plotted on top of the X-ray data points. The plot is divided into six different areas which correspondto six different X-ray states of Cygnus X-3 (see Fig. 2.6, same coloring schemeused). The corresponding X-ray state spectra are plotted in the upper part ofthe figure, each above its own area. The division between the hypersoft stateand the FSXR state is labeled the jet line. When Cygnus X-3 crosses this linefrom the hypersoft state, it produces a major flare. However, no flares havebeen observed when the system crosses the line in the other direction.

44

Results

line in the original HID, but the behavior of Cygnus X-3 across this line is

observed to be the opposite compared to other XRBs/microquasars: when

Cygnus X-3 crosses the jet line from right to left, radio jet ejections are not

observed and the source just quenches into the hypersoft state. However,

when the source exits the hypersoft state a major radio flare is always

observed. In addition, when Cygnus X-3 crosses the jet line, i.e. moving

into or out of the hypersoft state, a burst of γ-ray emission is detected (see

Section 2.3.2).

However, major flares differ depending on how strong the accompany-

ing hard X-ray flare is (Publication I). Variations can also occur within

the same flaring episode. After reaching the peak radio flux density of

the major flare, the X-ray spectra evolve from soft to hard (Szostek and

Zdziarski 2008, Publication I) with a rising non-thermal tail in the X-

ray spectra (i.e. FSXR–FIM states) until the tail exhibits a cutoff (FHXR

state). However, the FHXR and FIM states can occur in the opposite order

after the peak of the major flare.

4.2 Low-frequency QPOs and the jet

In Publication II, the search method for QPOs described in Section 3.2.2

was used to analyze the whole RXTE archive (at the time of publication)

including all X-ray spectral states to search for low-frequency QPOs (< 0.1

Hz). This extensive search resulted in two detections above the 99.9%

confidence level, further reinforcing the short-lived and sporadic nature

of these events. In addition, both of these detections as well as the pre-

vious ones in van der Klis and Jansen (1985) and Rao et al. (1991) are

associated to a certain extent with major radio flaring events (Publication

II;Publication VI). This result brings forth scenarios where the jet or some

structure in the jet is responsible for modulating the underlying emission

or causing the QPOs itself and thus broadening our notion of where the

QPOs originate.

4.3 Spectral components during an outburst

In Publication III PCA was performed on the X-ray spectra from RXTE,

INTEGRAL and Swift during a major flare ejection event in May–July

2006. The analysis showed that there are two main variability compo-

45

Results

54500 54600 54700 54800 54900 550000.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

MJD [days]

BA

T d

aily

flu

x (1

5−50

keV

) [c

ount

scm

−2s−1

]

Figure 4.2. The Swift/BAT ligthcurve with AGILE coverage (shaded grey bars) with theGeV detections (blue arrows). Note the occurrence of γ-ray detections in ev-ery local minimum of the hard X-ray lightcurve.

nents in play, i.e. two principal components accounted for almost all the

variability in the X-ray lightcurves. According to the spectral shape of

these components and spectral fits to the original data, the most proba-

ble emission components corresponding to the principal components are

inverse-Compton scattering and bremsstrahlung. In addition, during

the major flare ejection a plausible scenario of bremsstrahlung photons

inverse-Compton scattering off of relativistic electrons in the jet is sup-

ported by the analysis.

4.4 The link between γ-ray emission and spectral states

The γ-ray emission is associated with specific times characterized by

low hard X-ray flux, low radio flux and high soft X-ray flux. A strong

anti-correlation exists between the γ-ray emission and the hard X-ray

emission, and it appears that every local minimum in the hard X-ray

lightcurve (Swift/BAT count rate <0.02 counts cm−2 s−1) corresponds to a

time when γ-ray emission is detected from the system (Fig. 4.2). Likewise,

during the periods of γ-ray emission the RXTE/ASM count rate exceeds

the transitional value of 3 counts/s (Szostek and Zdziarski 2008). Thus,

the X-ray/radio state that corresponds to the above correlations is the hy-

persoft state. The γ-ray flares are observed during the declining phase as

well as the rising phase to/from the hypersoft state (Publication IV). The

hypersoft state is almost always associated with a major radio flare and

therefore a jet ejection event. When the source leaves the hypersoft state

46

Results

a major radio flare with a radio flux density of several Janskys is detected.

This ties in the γ-ray emission with the jet and it is very plausible that

the source of the γ-rays is the jet, as no other component in microquasars

is known to be responsible for generating copious numbers of relativistic

particles. Thus, the following sequence of events always takes place dur-

ing a major radio flare event: the source transits into the hypersoft state

producing γ-rays, and then remains in the hypersoft state with little or

no radio nor hard X-ray emission for approximately ten days up to one

month until it again flares in γ-rays; a few days after the appearance of

the γ-ray emission, the source abruptly brightens in the radio by several

orders of magnitude, signaling a jet ejection event, in conjunction with

the recovery of the hard X-ray emission. Studying this unique sequence

of events opens up a new avenue to exploring how the accretion disk, the

corona, and the jet are inter-connected.

47

Results

48

5. Discussion

Despite having been studied for 46 years at all wavelengths throughout

the electromagnetic spectrum, Cygnus X-3 still has a few surprises up

its sleeves as this thesis shows. With the advent of high-precision space-

based γ-ray observations we have seen that microquasars do indeed live

up to their multiwavelength nature as most recently manifested by the

detection of γ-ray emission from Cygnus X-3, but it could be that to truly

fulfill this criterion special conditions need to be met in order to extend

up to the GeV level. In addition, despite the heavy damping of the X-ray

timing signals a few low-frequency QPOs are detected which are interest-

ingly coincident with the major flaring episodes. The soft X-ray emission

from Cygnus X-3 could represent the characteristics of bremsstrahlung

emission, which is a relatively rare emission process in the microquasar

scenario. But do these findings present a more unified picture of Cygnus

X-3 or do they merely offer more mysteries to be solved by further re-

search? As necessary when regarding the essence of science, it is a bit of

both.

5.1 Implications

A key element in the observations of all the above-mentioned “surprises”

is to look at Cygnus X-3 at the right time – the “right” time being re-

stricted to a major radio flare, the messenger of a jet ejection event. Thus,

it appears that the jet is the culprit, or at least partly responsible for the

unique properties of Cygnus X-3. The order of the events during a major

flaring episode is as follows (Fig. 5.1):

I: Decline to radio quenched i.e. hypersoft state with γ-ray flares (though

the latter are not always present).

49

Discussion

Figure 5.1. Multiwavelength campaign from Piano et al. (2012) including X-ray monitor-ing from RXTE/ASM and Swift/BAT in the upper panel, and radio monitor-ing from AMI-LA and RATAN-600 and γ-ray monitoring from AGILE-GRIDin the lower panel. Roman numerals mark the sequence of events describedin the text.

50

Discussion

II: Hypersoft state characterized by low/non-existent levels of radio/hard

X-ray emission. This stage lasts from about ten days up to a month.

III: Rapid rise of radio emission with γ-ray flares accompanied with the

return of hard X-ray emission. Jet appears in the radio imaging.

IV: After the radio peak the X-ray spectrum hardens and the radio de-

clines or continues to produce more flares. QPOs are detected in a few

major flaring episodes at this stage.

V: Eventually the flaring episode powers down to a quiescent state where

Cygnus X-3 can remain for years until the next outburst.

What is remarkable about Cygnus X-3 is that the sequence of events

described above appears to operate backwards when compared to other

XRBs/microquasars. No major radio flares have been observed when

Cygnus X-3 transitions to its softest state which is normally the case for

other XRBs/microquasars. However, γ-ray flares are observed during this

time and thus they might convey information on the birth of the jet. What

remains to be solved in this scenario is why the radio emission lags the

γ-ray emission by 10–30 days.

The correlation between the radio and hard X-ray emission has been

established in McCollough et al. (1999), where positive correlation was

found during radio flaring (including the radio quenched/hypersoft state)

and negative correlation during the quiescent state. Publication I showed

that the X-ray intensity increases by a factor of 2–3 when the source tran-

sits to flaring states. The dominant emission process in the hard X-rays

is Comptonization as demonstrated in Publication I and Publication III,

and thus it seems plausible to assume that the jet is producing the hot

population of electrons required for Comptonizing the seed photons.

As all the jet launching mechanisms are dependent upon the existence

of an accretion disk, one of the most natural places for the soft seed pho-

ton population is from the accretion disk. However, multiple soft seed

population are possible and are hinted at in Publication III, where a

bremsstrahlung component produced the best match for producing part of

the soft X-ray emission (together with Comptonization). In any case a hot,

thermal component was needed to fit the spectra and produce the spectral

evolution of the principal components. Similarly, hot thermal components

51

Discussion

have been found to be necessary in order to fit well the intermediate X-

ray spectra of microquasars GRS 1915+105 (Titarchuk and Seifina 2009;

Mineo et al. 2012) and SS 433 (Seifina and Titarchuk 2010), and LMXRBs

GS 2000+25, GS 1124−68 and XTE J1550−564 (Zycki et al. 2001). This

implies that perhaps multiple soft photon populations are needed to fit

the spectra of other microquasars/XRBs as well. On the other hand, these

systems do not contain a massive stellar wind emanating from the com-

panion star as in the case of Cygnus X-3. One option could be that an

accretion disk wind in these other systems could mimic the effect of a

stellar wind. Likewise, it would be possible that a massive disk wind is

mimicking the WR-phenomena for Cygnus X-3, but so far no such model

exists.

Still another characteristic unique to Cygnus X-3 is that the jet axis ap-

pear to be aligned or close to the line-of-sight. Mioduszewski et al. (2001)

found that the jets are inclined 14 degrees to the line of sight and pre-

cessing. Reinterpreting the archival Very Long Baseline Interferometry

images, Miller-Jones et al. (2009) showed that the jets can change their

position angle by 180 degrees with time which is in support of the preces-

sion and the jet axis lying close to the line-of-sight. Based on modeling

the γ-ray emission and its orbital modulation Dubus et al. (2010) arrived

at jet parameters that lie close to the line-of-sight. These studies show

that Cygnus X-3 is more likely a “microblazar” than a microquasar. This

notion could help in understanding the brightness in the radio during ma-

jor radio flares as moderate beaming likely takes place (Miller-Jones et al.

2004; Lindfors et al. 2007). The “blazarness” of Cygnus X-3 might also ex-

plain the origin of the low-frequency QPOs during major flaring episodes.

The QPOs could be caused by some structure in the jet emitting X-rays,

e.g. a shock. Rani et al. (2010) have shown that low-frequency QPOs

could occur in blazars in a turbulent region behind a shock where domi-

nant eddies with turnover times corresponding to the QPO periods could

explain the short-lived, quasi-periodic fluctuation. On the other hand,

Vilhu and Hannikainen (2013) discussed a scenario where the jet forms

clumps which end up trailing the jet and are advected together with the

stellar wind to larger distances. This trail of clumps could cause flick-

ering of the underlaying X-ray emission and thus cause quasi-periodic

fluctuation. An interesting aspect of this model is that the projection of

this clumpy trail onto the orbital plane places it in the phase interval

0.3–0.4, exactly the same phase interval where the low-frequency QPOs

52

Discussion

were observed in Publication II. This interval also exhibits an increase

in the overall X-ray variability predominantly in the harder X-ray states

(FIM/FHXR) observed during radio flaring. These states can be attributed

to times when jet ejection is observed while a softer X-ray state (FSXR)

might indicate a failed jet (Publication I). Whether or not the clumpy trail

turns out to be a plausible model, it nevertheless gives more credit to the

jet origin of the low-frequency QPOs.

The nature of the compact object in Cygnus X-3 has been debated ever

since its discovery with proponents in favor of a neutron star or a black

hole as the compact object. The galactic matter along the line-of-sight ab-

sorbs optical wavelengths and thus little is known about the inclination

and the mass function of Cygnus X-3. With the help of radial velocity mea-

surements of X-ray emission lines a separate case can be drawn for neu-

tron star or black hole alternatives (Publication V) with high inclination

(∼60 degrees) for the neutron star case and low inclination (∼30 degrees)

for the black hole case. Low inclination is supported by studying the or-

bital modulation with infrared (Publication V) and X-ray data (Zdziarski

et al. 2012a). Low inclination also fits well with the low inclination of

the jet assuming that the accretion disk is mostly aligned with the orbital

plane and the jet axis is mostly perpendicular to that. The mass estimates

for the black hole range between 2.8–8.0 M� (Publication V) and 1.3–4.5

M� (Zdziarski et al. 2013). While the low end of these ranges overlap the

allowed neutron star masses the black hole case is further strengthened

with the similarity of the X-ray spectral states of Cygnus X-3 with GRS

1915+105 and XTE J1550−564 (Szostek et al. 2008; Hjalmarsdotter et al.

2009). Also, a model-independent study by Vrtilek and Boroson (2013)

showed that Cygnus X-3 shares space between GRS 1915+105 -like black

hole systems and pulsars in the color-color-intensity volume. However, the

hallmark of pulsars – regular X-ray pulsations – have not been observed

from Cygnus X-3 and also the X-ray spectra of pulsars are rather different

to those observed from Cygnus X-3. Thus, the most likely compact object

is a low-mass black hole. Interestingly, 2–5 M� compact objects are not

found to abound in the Universe (Belczynski et al. 2012, and references

therein) and this compact object mass region is often dubbed the “mass

gap”. If indeed the low-mass black hole nature is confirmed in the future

this adds yet another unique aspect to Cygnus X-3.

A note should be made about possible neutrino emission from Cygnus

X-3. It has been estimated by Distefano et al. (2002) that Cygnus X-3

53

Discussion

should be the brightest neutrino emitter in the Galaxy, in the context

of a hadronic jet, where protons colliding with synchrotron photons pro-

duce neutrinos. In this scenario the neutrino flux would precede the γ-ray

flares, and thus the above-mentioned order of events in the major flaring

episode becomes important. If a neutrino flux is observed from Cygnus

X-3, it certainly would help assessing the composition of the jet and help

narrowing down the birth of the jet. However, neutrinos have not been

observed so far from Cygnus X-3 (IceCube Collaboration et al. 2013). On

the other hand, this gives more credibility to the leptonic composition of

the jet.

5.2 Concluding remarks

To conclude, it is evident that the jet plays a major role in the nature of

Cygnus X-3. By implementing a careful analysis of the X-ray data in the

spectral, timing and spectrotiming domains with supporting observations

across the electromagnetic spectrum, it has been demonstrated in this

thesis that new features can be found even in a source studied for such a

long time. These features include the softest X-ray state, i.e. the hypersoft

state, when the source exhibits quenching in the radio and hard X-rays.

When Cygnus X-3 descends into the hypersoft state γ-ray flares are ob-

served possibly indicating the birth of the jet. Likewise when Cygnus X-3

ascends from the hypersoft state γ-ray flares are observed again followed

by a major radio flare, the jet becoming visible in the radio images, and

the recovery of the hard X-rays. What remains to be solved is the overall

model to explain this behavior: the order of these events which appear to

occur in the inverse sense when compared to other microquasars/XRBs,

the nature of the hypersoft state and the mechanisms of the radio and

hard X-ray quenching and the mechanism producing the γ-ray emission.

In this thesis it has been debated that the link between the jet and the

hard X-rays is in the form of Comptonization: the jet producing the pop-

ulation of hot electrons Compton upscatters the seed photons thereby ex-

plaining the radio/hard X-ray correlation. By utilizing PCA to study the

time evolution of X-ray spectra it has been shown in this thesis that the

seed photon population can arise from two distinct populations, one colder

(presumably from the accretion disk) and one hotter (presumably from a

hot, thermal plasma around the compact object). The advantages of rul-

ing out spectral models with this method should be enticing to the X-ray

54

Discussion

community. Still more effects caused by the jet could be the sporadic de-

tections of low-frequency QPOs coinciding with the aftermath of the jet

ejection. While the origin of these QPOs remains to be investigated it is

interesting to note that they occur in the orbital phase when we should

observe the jet. Thus, it might be possible to study the connection be-

tween the jet and the interstellar medium close to Cygnus X-3 with the

aid of the underlying X-ray emission. While Cygnus X-3 is a very complex

source it is also very rewarding in that it produces one interesting phe-

nomenon after the other. Although a complete model of the source is yet

to be seen, hopefully this work has moved us a notch towards complete

understanding of this enigmatic microquasar.

55

Discussion

56

Bibliography

J. Aleksic, L. A. Antonelli, P. Antoranz, M. Backes, C. Baixeras, J. A. Bar-rio, D. Bastieri, J. Becerra González, W. Bednarek, A. Berdyugin, K. Berger,E. Bernardini, A. Biland, O. Blanch, R. K. Bock, A. Boller, G. Bonnoli, P. Bor-das, D. Borla Tridon, V. Bosch-Ramon, D. Bose, I. Braun, T. Bretz, D. Britzger,M. Camara, E. Carmona, A. Carosi, P. Colin, J. L. Contreras, J. Cortina, M. T.Costado, S. Covino, F. Dazzi, A. De Angelis, E. De Cea del Pozo, B. De Lotto,M. De Maria, F. De Sabata, C. Delgado Mendez, M. Doert, A. Domínguez,D. Dominis Prester, D. Dorner, M. Doro, D. Elsaesser, M. Errando, D. Ferenc,M. V. Fonseca, L. Font, R. J. García López, M. Garczarczyk, M. Gaug, N. Godi-novic, F. Göebel, D. Hadasch, A. Herrero, D. Hildebrand, D. Höhne-Mönch,J. Hose, D. Hrupec, C. C. Hsu, T. Jogler, S. Klepser, T. Krähenbühl, D. Kranich,A. La Barbera, A. Laille, E. Leonardo, E. Lindfors, S. Lombardi, F. Longo,M. López, E. Lorenz, P. Majumdar, G. Maneva, N. Mankuzhiyil, K. Mannheim,L. Maraschi, M. Mariotti, M. Martínez, D. Mazin, M. Meucci, J. M. Miranda,R. Mirzoyan, H. Miyamoto, J. Moldón, M. Moles, A. Moralejo, D. Nieto, K. Nils-son, J. Ninkovic, R. Orito, I. Oya, S. Paiano, R. Paoletti, J. M. Paredes, S. Par-tini, M. Pasanen, D. Pascoli, F. Pauss, R. G. Pegna, M. A. Perez-Torres, M. Per-sic, L. Peruzzo, F. Prada, E. Prandini, N. Puchades, I. Puljak, I. Reichardt,W. Rhode, M. Ribó, J. Rico, M. Rissi, S. Rügamer, A. Saggion, K. Saito, T. Y.Saito, M. Salvati, M. Sánchez-Conde, K. Satalecka, V. Scalzotto, V. Scapin,C. Schultz, T. Schweizer, M. Shayduk, S. N. Shore, A. Sierpowska-Bartosik,A. Sillanpää, J. Sitarek, D. Sobczynska, F. Spanier, S. Spiro, A. Stamerra,B. Steinke, J. C. Struebig, T. Suric, L. Takalo, F. Tavecchio, P. Temnikov,T. Terzic, D. Tescaro, M. Teshima, D. F. Torres, H. Vankov, R. M. Wagner,Q. Weitzel, V. Zabalza, F. Zandanel, R. Zanin, and MAGIC Collaboration.Magic Constraints on γ-ray Emission from Cygnus X-3. ApJ, 721:843–855,September 2010.

M. Axelsson, S. Larsson, and L. Hjalmarsdotter. The aperiodic broad-band X-rayvariability of Cygnus X-3. MNRAS, 394:1544–1550, April 2009.

S. A. Balbus and J. F. Hawley. Instability, turbulence, and enhanced transportin accretion disks. Reviews of Modern Physics, 70:1–53, January 1998.

G. Barbiellini, G. Fedel, F. Liello, F. Longo, C. Pontoni, M. Prest, M. Tavani,and E. Vallazza. The AGILE silicon tracker: testbeam results of the prototypesilicon detector. Nuclear Instruments and Methods in Physics Research A, 490:146–158, September 2002.

K. Belczynski, G. Wiktorowicz, C. L. Fryer, D. E. Holz, and V. Kalogera. Missing

57

Bibliography

black holes unveil the supernova explosion mechanism. ApJ, 757:91, Septem-ber 2012.

T. Belloni, J. Homan, P. Casella, M. van der Klis, E. Nespoli, W. H. G. Lewin,J. M. Miller, and M. Méndez. The evolution of the timing properties of theblack-hole transient GX 339-4 during its 2002/2003 outburst. A&A, 440:207–222, September 2005.

S. Benlloch, J. Wilms, R. Edelson, T. Yaqoob, and R. Staubert. Quasi-periodicoscillation in Seyfert galaxies: significance levels. The case of Markarian 766.ApJ, 562:L121–L124, December 2001.

P. R. Bevington and D. K. Robinson. Data reduction and error analysis for thephysical sciences. McGraw-Hill, 1992.

C. L. Bhat, M. L. Sapru, and H. Razdan. Estimate from Gulmarg of PeV photonflux from Cygnus X-3 and its relevance. ApJ, 306:587–598, July 1986.

B. Binder, B. F. Williams, M. Eracleous, M. R. Garcia, S. F. Anderson, and T. J.Gaetz. A deep Chandra observation of the Wolf-Rayet + black hole binary NGC300 X-1. ApJ, 742:128, December 2011.

R. D. Blandford. Accretion disc electrodynamics - A model for double radiosources. MNRAS, 176:465–481, September 1976.

R. D. Blandford and D. G. Payne. Hydromagnetic flows from accretion discs andthe production of radio jets. MNRAS, 199:883–903, June 1982.

R. D. Blandford and R. L. Znajek. Electromagnetic extraction of energy fromKerr black holes. MNRAS, 179:433–456, May 1977.

A. Bulgarelli, M. Tavani, A. W. Chen, Y. Evangelista, M. Trifoglio, F. Gianotti,G. Piano, S. Sabatini, E. Striani, G. Pooley, S. Trushkin, N. A. Nizhelskij,M. McCollough, K. I. I. Koljonen, D. Hannikainen, A. Lähteenmäki, J. Tammi,N. Lavonen, D. Steeghs, A. Aboudan, A. Argan, G. Barbiellini, R. Campana,P. Caraveo, P. W. Cattaneo, V. Cocco, T. Contessi, E. Costa, F. D’Ammando,E. Del Monte, G. de Paris, G. Di Cocco, I. Donnarumma, M. Feroci, M. Fiorini,F. Fuschino, M. Galli, A. Giuliani, M. Giusti, C. Labanti, I. Lapshov, F. Laz-zarotto, P. Lipari, F. Longo, M. Marisaldi, S. Mereghetti, E. Morelli, E. Moretti,A. Morselli, L. Pacciani, A. Pellizzoni, F. Perotti, P. Picozza, M. Pilia, M. Prest,G. Pucella, M. Rapisarda, A. Rappoldi, A. Rubini, P. Soffitta, A. Trois, E. Val-lazza, S. Vercellone, V. Vittorini, A. Zambra, D. Zanello, P. Giommi, C. Pit-tori, F. Verrecchia, P. Santolamazza, F. Lucarelli, S. Colafrancesco, and L. Sa-lotti. AGILE detection of Cygnus X-3 γ-ray active states during the periodmid-2009/mid-2010. A&A, 538:A63, February 2012.

T. Bulik, K. Belczynski, and A. Prestwich. IC10 X-1/NGC300 X-1: the very im-mediate progenitors of BH-BH binaries. ApJ, 730:140, April 2011.

S. Carpano, A. M. T. Pollock, A. Prestwich, P. Crowther, J. Wilms, L. Yungelson,and M. Ehle. A 33 hour period for the Wolf-Rayet/black hole X-ray binarycandidate NGC 300 X-1. A&A, 466:L17–L20, May 2007.

M. Choudhury, A. R. Rao, S. V. Vadawale, A. K. Jain, and N. S. Singh. Binary cor-rected X-ray light curve of Cygnus X-3: implications for the timing propertiesof the compact binary system. A&A, 420:665–671, June 2004.

58

Bibliography

S. Corbel, G. Dubus, J. A. Tomsick, A. Szostek, R. H. D. Corbet, J. C. A. Miller-Jones, J. L. Richards, G. Pooley, S. Trushkin, R. Dubois, A. B. Hill, M. Kerr,W. Max-Moerbeck, A. C. S. Readhead, A. Bodaghee, V. Tudose, D. Parent,J. Wilms, and K. Pottschmidt. A giant radio flare from Cygnus X-3 with asso-ciated γ-ray emission. MNRAS, 421:2947–2955, April 2012.

P. A. Crowther, R. Barnard, S. Carpano, J. S. Clark, V. S. Dhillon, and A. M. T.Pollock. NGC 300 X-1 is a Wolf-Rayet/black hole binary. MNRAS, 403:L41–L45, March 2010.

S. Danaher, D. J. Fegan, N. A. Porter, and T. C. Weekes. Gamma-ray observationsof Cygnus X-3 at energies of 10 to the 12th eV. Nat, 289:568–569, February1981.

G. Di Cocco, E. Caroli, E. Celesti, L. Foschini, F. Gianotti, C. Labanti,G. Malaguti, A. Mauri, E. Rossi, F. Schiavone, A. Spizzichino, J. B. Stephen,A. Traci, and M. Trifoglio. IBIS/PICsIT in-flight performances. A&A, 411:L189–L195, November 2003.

J. M. Dickey. A new distance to Cygnus X-3. ApJ, 273:L71–L73, October 1983.

C. Distefano, D. Guetta, E. Waxman, and A. Levinson. Neutrino flux predictionsfor known galactic microquasars. ApJ, 575:378–383, August 2002.

G. Dubus, B. Cerutti, and G. Henri. The relativistic jet of Cygnus X-3 in gamma-rays. MNRAS, 404:L55–L59, May 2010.

R. P. Fender, M. M. Hanson, and G. G. Pooley. Infrared spectroscopic variabilityof Cygnus X-3 in outburst and quiescence. MNRAS, 308:473–484, September1999.

R. P. Fender, T. M. Belloni, and E. Gallo. Towards a unified model for black holeX-ray binary jets. MNRAS, 355:1105–1118, December 2004.

R. P. Fender, J. Homan, and T. M. Belloni. Jets from black hole X-ray binaries:testing, refining and extending empirical models for the coupling to X-rays.MNRAS, 396:1370–1382, July 2009.

Fermi LAT Collaboration, A. A. Abdo, M. Ackermann, M. Ajello, M. Axelsson,L. Baldini, J. Ballet, G. Barbiellini, D. Bastieri, B. M. Baughman, K. Bech-tol, R. Bellazzini, B. Berenji, R. D. Blandford, E. D. Bloom, E. Bonamente,A. W. Borgland, A. Brez, M. Brigida, P. Bruel, T. H. Burnett, S. Buson, G. A.Caliandro, R. A. Cameron, P. A. Caraveo, J. M. Casandjian, C. Cecchi, Ö. Çe-lik, S. Chaty, C. C. Cheung, J. Chiang, S. Ciprini, R. Claus, J. Cohen-Tanugi,L. R. Cominsky, J. Conrad, S. Corbel, R. Corbet, C. D. Dermer, F. de Palma,S. W. Digel, E. do Couto e Silva, P. S. Drell, R. Dubois, G. Dubus, D. Dumora,C. Farnier, C. Favuzzi, S. J. Fegan, W. B. Focke, P. Fortin, M. Frailis, P. Fusco,F. Gargano, N. Gehrels, S. Germani, G. Giavitto, B. Giebels, N. Giglietto,F. Giordano, T. Glanzman, G. Godfrey, I. A. Grenier, M.-H. Grondin, J. E.Grove, L. Guillemot, S. Guiriec, Y. Hanabata, A. K. Harding, M. Hayashida,E. Hays, A. B. Hill, L. Hjalmarsdotter, D. Horan, R. E. Hughes, M. S. Jack-son, G. Jóhannesson, A. S. Johnson, T. J. Johnson, W. N. Johnson, T. Kamae,H. Katagiri, N. Kawai, M. Kerr, J. Knödlseder, M. L. Kocian, E. Koerding,M. Kuss, J. Lande, J. Latronico, M. Lemoine-Goumard, F. Longo, F. Loparco,B. Lott, M. N. Lovellette, P. Lubrano, G. M. Madejski, A. Makeev, L. Marchand,

59

Bibliography

M. Marelli, W. Max-Moerbeck, M. N. Mazziotta, N. McColl, J. E. McEnery,C. Meurer, P. F. Michelson, S. Migliari, W. Mitthumsiri, T. Mizuno, C. Monte,M. E. Monzani, A. Morselli, I. V. Moskalenko, S. Murgia, P. L. Nolan, J. P. Nor-ris, E. Nuss, T. Ohsugi, N. Omodei, R. A. Ong, J. F. Ormes, D. Paneque, D. Par-ent, V. Pelassa, M. Pepe, M. Pesce-Rollins, F. Piron, G. Pooley, T. A. Porter,K. Pottschmidt, S. Rainò, R. Rando, P. S. Ray, M. Razzano, N. Rea, A. Read-head, A. Reimer, O. Reimer, J. L. Richards, L. S. Rochester, J. Rodriguez, A. Y.Rodriguez, R. W. Romani, F. Ryde, H. F.-W. Sadrozinski, A. Sander, P. M. SazParkinson, C. Sgrò, E. J. Siskind, D. A. Smith, P. D. Smith, P. Spinelli, J.-L.Starck, M. Stevenson, M. S. Strickman, D. J. Suson, H. Takahashi, T. Tanaka,J. B. Thayer, D. J. Thompson, L. Tibaldo, J. A. Tomsick, D. F. Torres, G. Tosti,A. Tramacere, Y. Uchiyama, T. L. Usher, V. Vasileiou, N. Vilchez, V. Vitale, A. P.Waite, P. Wang, J. Wilms, B. L. Winer, K. W. Wood, T. Ylinen, and M. Ziegler.Modulated high-energy gamma-ray emission from the microquasar Cygnus X-3. Science, 326:1512–1516, December 2009.

M. Feroci, E. Costa, P. Soffitta, E. Del Monte, G. di Persio, I. Donnarumma,Y. Evangelista, M. Frutti, I. Lapshov, F. Lazzarotto, M. Mastropietro,E. Morelli, L. Pacciani, G. Porrovecchio, M. Rapisarda, A. Rubini, M. Tavani,and A. Argan. SuperAGILE: the hard X-ray imager for the AGILE space mis-sion. Nuclear Instruments and Methods in Physics Research A, 581:728–754,November 2007.

N. Gehrels, G. Chincarini, P. Giommi, K. O. Mason, J. A. Nousek, A. A. Wells,N. E. White, S. D. Barthelmy, D. N. Burrows, L. R. Cominsky, K. C. Hurley,F. E. Marshall, P. Mészáros, P. W. A. Roming, L. Angelini, L. M. Barbier, T. Bel-loni, S. Campana, P. A. Caraveo, M. M. Chester, O. Citterio, T. L. Cline, M. S.Cropper, J. R. Cummings, A. J. Dean, E. D. Feigelson, E. E. Fenimore, D. A.Frail, A. S. Fruchter, G. P. Garmire, K. Gendreau, G. Ghisellini, J. Greiner,J. E. Hill, S. D. Hunsberger, H. A. Krimm, S. R. Kulkarni, P. Kumar, F. Le-brun, N. M. Lloyd-Ronning, C. B. Markwardt, B. J. Mattson, R. F. Mushotzky,J. P. Norris, J. Osborne, B. Paczynski, D. M. Palmer, H.-S. Park, A. M. Parsons,J. Paul, M. J. Rees, C. S. Reynolds, J. E. Rhoads, T. P. Sasseen, B. E. Schaefer,A. T. Short, A. P. Smale, I. A. Smith, L. Stella, G. Tagliaferri, T. Takahashi,M. Tashiro, L. K. Townsley, J. Tueller, M. J. L. Turner, M. Vietri, W. Voges,M. J. Ward, R. Willingale, F. M. Zerbi, and W. W. Zhang. The Swift Gamma-ray Burst Mission. ApJ, 611:1005–1020, August 2004.

R. Giacconi, P. Gorenstein, H. Gursky, and J. R. Waters. An X-ray survey of theCygnus region. ApJ, 148:L119–L127, June 1967.

W. Hermsen, J. B. G. M. Bloemen, F. A. Jansen, K. Bennett, R. Buccheri, A. Mas-tichiadis, H. A. Mayer-Hasselwander, A. W. Strong, M. E. Oezel, and A. M. T.Pollock. High-energy gamma-ray and hard X-ray observations of Cyg X-3.A&A, 175:141–150, March 1987.

L. Hjalmarsdotter, A. A. Zdziarski, A. Szostek, and D. C. Hannikainen. Spectralvariability in Cygnus X-3. MNRAS, 392:251–263, January 2009.

IceCube Collaboration, R. Abbasi, Y. Abdou, M. Ackermann, J. Adams, J. A.Aguilar, M. Ahlers, D. Altmann, K. Andeen, J. Auffenberg, X. Bai, M. Baker,S. W. Barwick, V. Baum, R. Bay, K. Beattie, J. J. Beatty, S. Bechet, J. BeckerTjus, K.-H. Becker, M. Bell, M. L. Benabderrahmane, S. BenZvi, J. Berder-mann, P. Berghaus, D. Berley, E. Bernardini, D. Bertrand, D. Z. Besson,

60

Bibliography

D. Bindig, M. Bissok, E. Blaufuss, J. Blumenthal, D. J. Boersma, C. Bohm,D. Bose, S. Böser, O. Botner, L. Brayeur, A. M. Brown, R. Bruijn, J. Brunner,S. Buitink, M. Carson, J. Casey, M. Casier, D. Chirkin, B. Christy, F. Clever-mann, S. Cohen, D. F. Cowen, A. H. Cruz Silva, M. Danninger, J. Daughhetee,J. C. Davis, C. De Clercq, F. Descamps, P. Desiati, G. de Vries-Uiterweerd,T. DeYoung, J. C. Díaz-Vélez, J. Dreyer, J. P. Dumm, M. Dunkman, R. Eagan,J. Eisch, R. W. Ellsworth, O. Engdegård, S. Euler, P. A. Evenson, O. Fadiran,A. R. Fazely, A. Fedynitch, J. Feintzeig, T. Feusels, K. Filimonov, C. Finley,T. Fischer-Wasels, S. Flis, A. Franckowiak, R. Franke, K. Frantzen, T. Fuchs,T. K. Gaisser, J. Gallagher, L. Gerhardt, L. Gladstone, T. Glüsenkamp, A. Gold-schmidt, J. A. Goodman, D. Góra, D. Grant, A. Groß, S. Grullon, M. Gurt-ner, C. Ha, A. Haj Ismail, A. Hallgren, F. Halzen, K. Hanson, D. Heereman,P. Heimann, D. Heinen, K. Helbing, R. Hellauer, S. Hickford, G. C. Hill, K. D.Hoffman, R. Hoffmann, A. Homeier, K. Hoshina, W. Huelsnitz, P. O. Hulth,K. Hultqvist, S. Hussain, A. Ishihara, E. Jacobi, J. Jacobsen, G. S. Japaridze,O. Jlelati, A. Kappes, T. Karg, A. Karle, J. Kiryluk, F. Kislat, J. Kläs, S. R.Klein, J.-H. Köhne, G. Kohnen, H. Kolanoski, L. Köpke, C. Kopper, S. Kopper,D. J. Koskinen, M. Kowalski, M. Krasberg, G. Kroll, J. Kunnen, N. Kurahashi,T. Kuwabara, M. Labare, K. Laihem, H. Landsman, M. J. Larson, R. Lauer,M. Lesiak-Bzdak, J. Lünemann, J. Madsen, R. Maruyama, K. Mase, H. S. Ma-tis, F. McNally, K. Meagher, M. Merck, P. Mészáros, T. Meures, S. Miarecki,E. Middell, N. Milke, J. Miller, L. Mohrmann, T. Montaruli, R. Morse, S. M.Movit, R. Nahnhauer, U. Naumann, S. C. Nowicki, D. R. Nygren, A. Obertacke,S. Odrowski, A. Olivas, M. Olivo, A. O’Murchadha, S. Panknin, L. Paul, C. Pep-per, J. A. Pérez de los Heros, D. Pieloth, N. Pirk, J. Posselt, P. B. Price, G. T.Przybylski, L. Rädel, K. Rawlins, P. Redl, E. Resconi, W. Rhode, M. Ribordy,M. Richman, B. Riedel, J. P. Rodrigues, F. Rothmaier, C. Rott, T. Ruhe, B. Ruzy-bayev, D. Ryckbosch, S. M. Saba, T. Salameh, H.-G. Sander, M. Santander,S. Sarkar, K. Schatto, M. Scheel, F. Scheriau, T. Schmidt, M. Schmitz, S. Schoe-nen, S. Schöneberg, L. Schönherr, A. Schönwald, A. Schukraft, L. Schulte,O. Schulz, D. Seckel, S. H. Seo, Y. Sestayo, S. Seunarine, M. W. E. Smith,M. Soiron, D. Soldin, G. M. Spiczak, C. Spiering, M. Stamatikos, T. Stanev,A. Stasik, T. Stezelberger, R. G. Stokstad, A. Stößl, E. A. Strahler, R. Ström,G. W. Sullivan, H. Taavola, I. Taboada, A. Tamburro, S. Ter-Antonyan, S. Tilav,P. A. Toale, S. Toscano, M. Usner, D. van der Drift, N. van Eijndhoven, A. VanOverloop, J. van Santen, M. Vehring, M. Voge, C. Walck, T. Waldenmaier,M. Wallraff, M. Walter, R. Wasserman, Ch. Weaver, C. Wendt, S. Westerhoff,N. Whitehorn, K. Wiebe, C. H. Wiebusch, D. R. Williams, H. Wissing, M. Wolf,T. R. Wood, K. Woschnagg, C. Xu, D. L. Xu, X. W. Xu, J. P. Yanez, G. Yodh,S. Yoshida, P. Zarzhitsky, J. Ziemann, A. Zilles, and M. Zoll. Searches forhigh-energy neutrino emission in the Galaxy with the combined ICECUBE-AMANDA detector. ApJ, 763:33, January 2013.

S. Ichimaru. Bimodal behavior of accretion disks - Theory and application toCygnus X-1 transitions. ApJ, 214:840–855, June 1977.

K. Jahoda, J. H. Swank, A. B. Giles, M. J. Stark, T. Strohmayer, W. Zhang, andE. H. Morgan. In-orbit performance and calibration of the Rossi X-ray TimingExplorer (RXTE) Proportional Counter Array (PCA). In O. H. Siegmund andM. A. Gummin, editors, Society of Photo-Optical Instrumentation Engineers(SPIE) Conference Series, volume 2808 of Society of Photo-Optical Instrumen-tation Engineers (SPIE) Conference Series, pages 59–70, October 1996.

61

Bibliography

R. C. Lamb, C. P. Godfrey, W. A. Wheaton, and T. Tumer. Cygnus X-3 observedat photon energies above 500 GeV. Nat, 296:543–544, April 1982.

F. Lebrun, J. P. Leray, P. Lavocat, J. Crétolle, M. Arquès, C. Blondel, C. Bon-nin, A. Bouère, C. Cara, T. Chaleil, F. Daly, F. Desages, H. Dzitko, B. Horeau,P. Laurent, O. Limousin, F. Mathy, V. Mauguen, F. Meignier, F. Molinié,E. Poindron, M. Rouger, A. Sauvageon, and T. Tourrette. ISGRI: the INTE-GRAL Soft Gamma-Ray Imager. A&A, 411:L141–L148, November 2003.

A. M. Levine, H. Bradt, W. Cui, J. G. Jernigan, E. H. Morgan, R. Remillard, R. E.Shirey, and D. A. Smith. First results from the All-Sky Monitor on the RossiX-ray Timing Explorer. ApJ, 469:L33–L36, September 1996.

E. J. Lindfors, M. Türler, D. C. Hannikainen, G. Pooley, J. Tammi, S. A. Trushkin,and E. Valtaoja. Synchrotron flaring behaviour of Cygnus X-3 during theFebruary-March 1994 and September 2001 outbursts. A&A, 473:923–929, Oc-tober 2007.

Z. Ling, S. N. Zhang, and S. Tang. Determining the distance of Cyg X-3 with itsX-ray dust scattering halo. ApJ, 695:1111–1120, April 2009.

Q. Z. Liu, J. van Paradijs, and E. P. J. van den Heuvel. Catalogue of high-massX-ray binaries in the Galaxy (4th edition). A&A, 455:1165–1168, September2006.

Q. Z. Liu, J. van Paradijs, and E. P. J. van den Heuvel. A catalogue of low-massX-ray binaries in the Galaxy, LMC, and SMC (Fourth edition). A&A, 469:807–810, July 2007.

D. Lommen, L. Yungelson, E. van den Heuvel, G. Nelemans, and S. PortegiesZwart. Cygnus X-3 and the problem of the missing Wolf-Rayet X-ray binaries.A&A, 443:231–241, November 2005.

M. S. Longair. High Energy Astrophysics. Cambridge University Press, 2011.

R. V. E. Lovelace. Dynamo model of double radio sources. Nat, 262:649–652,August 1976.

N. Lund, C. Budtz-Jørgensen, N. J. Westergaard, S. Brandt, I. L. Rasmussen,A. Hornstrup, C. A. Oxborrow, J. Chenevez, P. A. Jensen, S. Laursen, K. H.Andersen, P. B. Mogensen, I. Rasmussen, K. Omø, S. M. Pedersen, J. Polny,H. Andersson, T. Andersson, V. Kämäräinen, O. Vilhu, J. Huovelin, S. Maisala,M. Morawski, G. Juchnikowski, E. Costa, M. Feroci, A. Rubini, M. Rapisarda,E. Morelli, V. Carassiti, F. Frontera, C. Pelliciari, G. Loffredo, S. MartínezNúñez, V. Reglero, T. Velasco, S. Larsson, R. Svensson, A. A. Zdziarski,A. Castro-Tirado, P. Attina, M. Goria, G. Giulianelli, F. Cordero, M. Rezazad,M. Schmidt, R. Carli, C. Gomez, P. L. Jensen, G. Sarri, A. Tiemon, A. Orr,R. Much, P. Kretschmar, and H. W. Schnopper. JEM-X: the X-ray monitoraboard INTEGRAL. A&A, 411:L231–L238, November 2003.

J. M. Mas-Hesse, A. Giménez, J. L. Culhane, C. Jamar, B. McBreen, J. Torra,R. Hudec, J. Fabregat, E. Meurs, J. P. Swings, M. A. Alcacera, A. Balado,R. Beiztegui, T. Belenguer, L. Bradley, M. D. Caballero, P. Cabo, J. M. Defise,E. Díaz, A. Domingo, F. Figueras, I. Figueroa, L. Hanlon, F. Hroch, V. Hudcova,T. García, B. Jordan, C. Jordi, P. Kretschmar, C. Laviada, M. March, E. Martín,E. Mazy, M. Menéndez, J. M. Mi, E. de Miguel, T. Muñoz, K. Nolan, R. Olmedo,

62

Bibliography

J. Y. Plesseria, J. Polcar, M. Reina, E. Renotte, P. Rochus, A. Sánchez, J. C. SanMartín, A. Smith, J. Soldan, P. Thomas, V. Timón, and D. Walton. OMC: anOptical Monitoring Camera for INTEGRAL. Instrument description and per-formance. A&A, 411:L261–L268, November 2003.

K. O. Mason, F. A. Cordova, and N. E. White. Simultaneous X-ray and infraredobservations of Cygnus X-3. ApJ, 309:700–706, October 1986.

J. E. McClintock and R. A. Remillard. Black hole binaries. In W. H. G. Lewin andM. van der Klis, editors, Compact stellar X-ray sources, pages 157–213. April2006.

M. L. McCollough, C. R. Robinson, S. N. Zhang, B. A. Harmon, R. M. Hjellming,E. B. Waltman, R. S. Foster, F. D. Ghigo, M. S. Briggs, G. N. Pendleton, andK. J. Johnston. Discovery of correlated behavior between the hard X-ray andthe radio bands in Cygnus X-3. ApJ, 517:951–955, June 1999.

M. L. McConnell, A. A. Zdziarski, K. Bennett, H. Bloemen, W. Collmar,W. Hermsen, L. Kuiper, W. Paciesas, B. F. Phlips, J. Poutanen, J. M. Ryan,V. Schönfelder, H. Steinle, and A. W. Strong. The soft gamma-ray spectralvariability of Cygnus X-1. ApJ, 572:984–995, June 2002.

S. Migliari, R. P. Fender, M. Rupen, P. G. Jonker, M. Klein-Wolt, R. M. Hjellming,and M. van der Klis. Disc-jet coupling in an atoll-type neutron star X-raybinary: 4U 1728-34 (GX 354-0). MNRAS, 342:L67–L71, July 2003.

J. C. A. Miller-Jones, K. M. Blundell, M. P. Rupen, A. J. Mioduszewski, P. Duffy,and A. J. Beasley. Time-sequenced multi-radio frequency observations ofCygnus X-3 in flare. ApJ, 600:368–389, January 2004.

J. C. A. Miller-Jones, C. M. Sakari, V. Dhawan, V. Tudose, R. P. Fender, Z. Paragi,and M. Garrett. The proper motion and changing jet morphology of Cygnus X-3. In Conway J. E., editor, The 8th International e-VLBI Workshop, volumeEXPReS09 of Proceedings of Science (PoS), page 17, June 2009.

T. Mineo, E. Massaro, A. D’Ai, F. Massa, M. Feroci, G. Ventura, P. Casella,C. Ferrigno, and T. Belloni. The complex behaviour of the microquasar GRS1915+105 in the ρ class observed with BeppoSAX. II. Time-resolved spectralanalysis. A&A, 537:A18, January 2012.

A. J. Mioduszewski, M. P. Rupen, R. M. Hjellming, G. G. Pooley, and E. B. Walt-man. A one-sided highly relativistic jet from Cygnus X-3. ApJ, 553:766–775,June 2001.

I. F. Mirabel and L. F. Rodríguez. A superluminal source in the Galaxy. Nat, 371:46–48, September 1994.

M. Mori, D. L. Bertsch, B. L. Dingus, J. A. Esposito, C. E. Fichtel, S. D. Hunter,G. Kanbach, D. A. Kniffen, Y. C. Lin, J. R. Mattox, H. A. Mayer-Hasselwander,P. F. Michelson, C. von Montigny, R. Mukherjee, P. L. Nolan, P. V. Ramana-murthy, E. Schneid, P. Sreekumar, and D. J. Thompson. Cygnus X-3 andEGRET Gamma-Ray Observations. ApJ, 476:842, February 1997.

R. Narayan and I. Yi. Advection-dominated accretion: a self-similar solution.ApJ, 428:L13–L16, June 1994.

63

Bibliography

M. A. Nowak. Toward a Unified View of Black-Hole High-Energy States. PASP,107:1207–1214, December 1995.

K. S. O’Flaherty, M. F. Cawley, D. J. Fegan, R. C. Lamb, M. J. Lang, D. A. Lewis,M. Punch, P. T. Reynolds, G. Vacanti, and T. C. Weekes. A search for TeVemission from Cygnus X-3. ApJ, 396:674–678, September 1992.

D. R. Parsignault, H. Gursky, E. M. Kellogg, T. Matilsky, S. Murray, E. Schreier,H. Tananbaum, R. Giacconi, and A. C. Brinkman. Observations of Cygnus X-3by Uhuru. Nature Physical Science, 239:123–125, October 1972.

G. Piano, M. Tavani, V. Vittorini, A. Trois, A. Giuliani, A. Bulgarelli, Y. Evange-lista, P. Coppi, E. Del Monte, S. Sabatini, E. Striani, I. Donnarumma, D. Han-nikainen, K. I. I. Koljonen, M. McCollough, G. Pooley, S. Trushkin, R. Zanin,G. Barbiellini, M. Cardillo, P. W. Cattaneo, A. W. Chen, S. Colafrancesco, M. Fe-roci, F. Fuschino, M. Giusti, F. Longo, A. Morselli, A. Pellizzoni, C. Pittori,G. Pucella, M. Rapisarda, A. Rappoldi, P. Soffitta, M. Trifoglio, S. Vercellone,and F. Verrecchia. The AGILE monitoring of Cygnus X-3: transient gamma-ray emission and spectral constraints. A&A, 545:A110, September 2012.

P. Predehl and J. H. M. M. Schmitt. X-raying the interstellar medium: ROSATobservations of dust scattering halos. A&A, 293:889–905, January 1995.

P. Predehl, V. Burwitz, F. Paerels, and J. Trümper. Chandra measurement ofthe geometrical distance to Cyg X-3 using its X-ray scattering halo. A&A, 357:L25–L28, May 2000.

P. Predehl, V. Burwitz, E. Costantini, and J. Trümper. Geometrical distance mea-surement of Cyg X-3 using its X-ray scattering halo. In H. Inoue and H. Ku-nieda, editors, New Century of X-ray Astronomy, volume 251 of AstronomicalSociety of the Pacific Conference Series, pages 46–49, 2001.

M. Prest, G. Barbiellini, G. Bordignon, G. Fedel, F. Liello, F. Longo, C. Pontoni,and E. Vallazza. The AGILE silicon tracker: an innovative /γ-ray instrumentfor space. Nuclear Instruments and Methods in Physics Research A, 501:280–287, March 2003.

A. H. Prestwich, R. Kilgard, P. A. Crowther, S. Carpano, A. M. T. Pollock,A. Zezas, S. H. Saar, T. P. Roberts, and M. J. Ward. The orbital period ofthe Wolf-Rayet binary IC 10 X-1: dynamic evidence that the compact object isa black hole. ApJ, 669:L21–L24, November 2007.

B. Punsly and F. V. Coroniti. Ergosphere-driven winds. ApJ, 354:583–615, May1990.

B. Rani, A. C. Gupta, U. C. Joshi, S. Ganesh, and P. J. Wiita. Quasi-periodic os-cillations of ∼15 minutes in the optical light curve of the BL Lac S5 0716+714.ApJ, 719:L153–L157, August 2010.

A. R. Rao, P. C. Agrawal, and R. K. Manchanda. Hard X-ray observations ofCygnus X-3. A&A, 241:127–130, January 1991.

R. C. Reis, J. M. Miller, and A. C. Fabian. Thermal emission from the stellar-mass black hole binary XTE J1118+480 in the low/hard state. MNRAS, 395:L52–L56, May 2009.

64

Bibliography

G. E. Romero, D. F. Torres, M. M. Kaufman Bernadó, and I. F. Mirabel. Hadronicgamma-ray emission from windy microquasars. A&A, 410:L1–L4, October2003.

R. E. Rothschild, P. R. Blanco, D. E. Gruber, W. A. Heindl, D. R. MacDonald, D. C.Marsden, M. R. Pelling, L. R. Wayne, and P. L. Hink. In-flight performanceof the High-Energy X-ray Timing Experiment on the Rossi X-ray Timing Ex-plorer. ApJ, 496:538–549, March 1998.

A. Rushton, J. C. A. Miller-Jones, R. Campana, Y. Evangelista, Z. Paragi,T. J. Maccarone, G. G. Pooley, V. Tudose, R. P. Fender, R. E. Spencer, andV. Dhawan. A weak compact jet in a soft state of Cygnus X-1. MNRAS, 419:3194–3199, February 2012.

G. B. Rybicki and A. P. Lightman. Radiative processes in astrophysics. Wiley-Interscience, 1979.

M. Samorski and W. Stamm. Detection of 2 × 1015 to 2 × 1016 eV gamma-raysfrom Cygnus X-3. ApJ, 268:L17–L21, May 1983.

E. Seifina and L. Titarchuk. On the nature of the compact object in SS 433:observational evidence of X-ray photon index saturation. ApJ, 722:586–604,October 2010.

N. I. Shakura and R. A. Sunyaev. Black holes in binary systems. Observationalappearance. A&A, 24:337–355, 1973.

J. M. Silverman and A. V. Filippenko. On IC 10 X-1, the most massive knownstellar-mass black hole. ApJ, 678:L17–L20, May 2008.

A. Szostek and A. A. Zdziarski. Effects of the stellar wind on X-ray spectra ofCygnus X-3. MNRAS, 386:593–607, May 2008.

A. Szostek, A. A. Zdziarski, and P. Lachowicz. RXTE observations of Cygnus X-3.In P. Kaaret, F. K. Lamb, and J. H. Swank, editors, X-ray Timing 2003: Rossiand Beyond, volume 714 of American Institute of Physics Conference Series,pages 113–115, July 2004.

A. Szostek, A. A. Zdziarski, and M. L. McCollough. A classification of the X-rayand radio states of Cyg X-3 and their long-term correlations. MNRAS, 388:1001–1010, August 2008.

H. Tananbaum, H. Gursky, E. Kellogg, R. Giacconi, and C. Jones. Observationof a Correlated X-Ray Transition in Cygnus X-1. ApJ, 177:L5–L10, October1972.

S. P. Tigelaar, R. P. Fender, R. P. J. Tilanus, E. Gallo, and G. G. Pooley. Quenchedmillimetre emission from Cygnus X-1 in a soft X-ray state. MNRAS, 352:1015–1018, August 2004.

J. Timmer and M. Koenig. On generating power law noise. A&A, 300:707–710,August 1995.

L. Titarchuk and E. Seifina. Discovery of photon index saturation in the blackhole binary GRS 1915+105. ApJ, 706:1463–1483, December 2009.

65

Bibliography

J. A. Tomsick, K. Yamaoka, S. Corbel, P. Kaaret, E. Kalemci, and S. Migliari.Truncation of the inner accretion disk around a black hole at low luminosity.ApJ, 707:L87–L91, December 2009.

S. A. Trushkin, N. N. Bursov, and N. A. Nizhelskij. Flaring variability of micro-quasars. In S. K. Chakrabarti and A. S. Majumdar, editors, American Insti-tute of Physics Conference Series, volume 1053 of American Institute of PhysicsConference Series, pages 219–224, October 2008.

V. Tudose, J. C. A. Miller-Jones, R. P. Fender, Z. Paragi, C. Sakari, A. Szostek,M. A. Garrett, V. Dhawan, A. Rushton, R. E. Spencer, and M. van der Klis.Probing the behaviour of the X-ray binary Cygnus X-3 with very long baselineradio interferometry. MNRAS, 401:890–900, January 2010.

P. Ubertini, F. Lebrun, G. Di Cocco, A. Bazzano, A. J. Bird, K. Broenstad, A. Gold-wurm, G. La Rosa, C. Labanti, P. Laurent, I. F. Mirabel, E. M. Quadrini,B. Ramsey, V. Reglero, L. Sabau, B. Sacco, R. Staubert, L. Vigroux, M. C. Weiss-kopf, and A. A. Zdziarski. IBIS: the Imager on-Board INTEGRAL. A&A, 411:L131–L139, November 2003.

M. van der Klis. Rapid X-ray Variability. In W. H. G. Lewin and M. van der Klis,editors, Compact stellar X-ray sources, pages 39–112. Cambridge UniversityPress, April 2006.

M. van der Klis. Similarities in neutron star and black hole accretion. ApJS, 92:511–519, June 1994.

M. van der Klis and F. A. Jansen. Transient quasi-periodic oscillations in theX-ray flux of Cygnus X-3. Nat, 313:768–771, February 1985.

M. H. van Kerkwijk, P. A. Charles, T. R. Geballe, D. L. King, G. K. Miley, L. A.Molnar, E. P. J. van den Heuvel, M. van der Klis, and J. van Paradijs. In-frared helium emission lines from Cygnus X-3 suggesting a Wolf-Rayet starcompanion. Nat, 355:703–705, February 1992.

G. Vedrenne, J.-P. Roques, V. Schönfelder, P. Mandrou, G. G. Lichti, A. von Kien-lin, B. Cordier, S. Schanne, J. Knödlseder, G. Skinner, P. Jean, F. Sanchez,P. Caraveo, B. Teegarden, P. von Ballmoos, L. Bouchet, P. Paul, J. Matteson,S. Boggs, C. Wunderer, P. Leleux, G. Weidenspointner, P. Durouchoux, R. Diehl,A. Strong, M. Cassé, M. A. Clair, and Y. André. SPI: the Spectrometer aboardINTEGRAL. A&A, 411:L63–L70, November 2003.

O. Vilhu and D. C. Hannikainen. Modeling the X-ray light curves of Cygnus X-3.Possible role of the jet. A&A, 550:A48, February 2013.

O. Vilhu, L. Hjalmarsdotter, A. A. Zdziarski, A. Paizis, M. L. McCollough,V. Beckmann, T. J.-L. Courvoisier, K. Ebisawa, P. Goldoni, P. Hakala, D. Han-nikainen, P. Kretschmar, and N. J. Westergaard. First INTEGRAL observa-tions of Cygnus X-3. A&A, 411:L405–L410, November 2003.

B. M. Vladimirsky, A. A. Stepanian, and V. P. Fomin. High-energy gamma-rayoutburst in the direction of the X-ray source Cyg X-3. In International CosmicRay Conference, volume 1 of International Cosmic Ray Conference, pages 456–460, 1973.

66

Bibliography

S. D. Vrtilek and B. S. Boroson. Model independent means of categorizing X-ray binaries - I. Colour-colour-intensity diagrams. MNRAS, 428:3693–3714,February 2013.

E. B. Waltman, F. D. Ghigo, K. J. Johnston, R. S. Foster, R. L. Fiedler, and J. H.Spencer. The evolution of outbursts in Cygnus X-3 at 2.25 and 8.3 GHz. AJ,110:290–299, July 1995.

E. B. Waltman, R. S. Foster, G. G. Pooley, R. P. Fender, and F. D. Ghigo. Quenchedradio emission in Cygnus X-3. Astronomical Journal, 112:2690–2698, Decem-ber 1996.

P. K. G. Williams, J. A. Tomsick, A. Bodaghee, G. C. Bower, G. G. Pooley,K. Pottschmidt, J. Rodriguez, J. Wilms, S. Migliari, and S. A. Trushkin. The2010 May Flaring Episode of Cygnus X-3 in Radio, X-rays, and γ-rays. ApJ,733:L20, June 2011.

R. Willingale, A. R. King, and K. A. Pounds. EXOSAT MEDA observations ofCygnus X-3. MNRAS, 215:295–314, July 1985.

A. A. Zdziarski, R. Misra, and M. Gierlinski. Compton scattering as the ex-planation of the peculiar X-ray properties of Cyg X-3. MNRAS, 402:767–775,February 2010.

A. A. Zdziarski, G. K. Skinner, G. G. Pooley, and P. Lubinski. X-ray variabilitypatterns and radio/X-ray correlations in Cyg X-1. MNRAS, 416:1324–1339,September 2011.

A. A. Zdziarski, C. Maitra, A. Frankowski, G. K. Skinner, and R. Misra. Energy-dependent orbital modulation of X-rays and constraints on emission of the jetin Cyg X-3. MNRAS, 426:1031–1042, October 2012a.

A. A. Zdziarski, M. Sikora, G. Dubus, F. Yuan, B. Cerutti, and A. Ogorzałek. Thegamma-ray emitting region of the jet in Cyg X-3. MNRAS, 421:2956–2968,April 2012b.

A. A. Zdziarski, J. Mikołajewska, and K. Belczynski. Cyg X-3: a low-mass blackhole or a neutron star. MNRAS, 429:L104–L108, February 2013.

J. T. L. Zwart, R. W. Barker, P. Biddulph, D. Bly, R. C. Boysen, A. R. Brown,C. Clementson, M. Crofts, T. L. Culverhouse, J. Czeres, R. J. Dace, M. L.Davies, R. D’Alessandro, P. Doherty, K. Duggan, J. A. Ely, M. Felvus, F. Feroz,W. Flynn, T. M. O. Franzen, J. Geisbüsch, R. Génova-Santos, K. J. B. Grainge,W. F. Grainger, D. Hammett, R. E. Hills, M. P. Hobson, C. M. Holler, N. Hurley-Walker, R. Jilley, M. E. Jones, T. Kaneko, R. Kneissl, K. Lancaster, A. N.Lasenby, P. J. Marshall, F. Newton, O. Norris, I. Northrop, D. M. Odell, G. Pe-tencin, J. C. Pober, G. G. Pooley, M. W. Pospieszalski, V. Quy, C. Rodríguez-Gonzálvez, R. D. E. Saunders, A. M. M. Scaife, J. Schofield, P. F. Scott, C. Shaw,T. W. Shimwell, H. Smith, A. C. Taylor, D. J. Titterington, M. Velic, E. M. Wal-dram, S. West, B. A. Wood, G. Yassin, and AMI Consortium. The ArcminuteMicrokelvin Imager. MNRAS, 391:1545–1558, December 2008.

P. T. Zycki, C. Done, and D. A. Smith. On the complex disc-corona interactionsin the soft spectral states of soft X-ray transients. MNRAS, 326:1367–1380,October 2001.

67

Bibliography

68

9HSTFMG*afcjga+

ISBN 978-952-60-5296-0 ISBN 978-952-60-5297-7 (pdf) ISSN-L 1799-4934 ISSN 1799-4934 ISSN 1799-4942 (pdf) Aalto University School of Electrical Engineering Department of Radio Science and Engineering www.aalto.fi

BUSINESS + ECONOMY ART + DESIGN + ARCHITECTURE SCIENCE + TECHNOLOGY CROSSOVER DOCTORAL DISSERTATIONS

Aalto-D

D 12

9/2

013

Extremely energetic radiation observed in the universe is attributed to the accretion of matter by compact objects: black holes and neutron stars. Microquasars, dubbed as such because of their similar phenomenological appearance to extragalactic quasars, are accreting compact objects with resolved, relativistic jets emanating from the central source. Cygnus X-3 is one of the most peculiar sources amongst microquasars. It is one of the most radio- and X-ray-luminous objects in our Galaxy, and thus a prime target for studying the emission mechanisms that arise during violent outbursts detected from the source throughout the electromagnetic spectrum. This thesis consists of constructing a hardness-intensity diagram from X-ray data with simultaneous radio observations from Cygnus X-3, searching for low-frequency quasi-periodic oscillations and using principal component analysis to distinguish the variability of emission components during outbursts.

Karri I. I. K

oljonen T

he multiw

avelength spectral and timing nature of the m

icroquasar Cygnus X

-3 A

alto U

nive

rsity

Department of Radio Science and Engineering



DOCTORAL DISSERTATIONS

the multiwavelength spectral and timing nature of the...

Documents